J. Dairy Sci. 88:4157-4164
© American Dairy Science Association, 2005.
Development and Application of Image Analysis to Quantify Calcium Lactate Crystals on the Surface of Smoked Cheddar Cheese*
P. Rajbhandari and
P. S. Kindstedt
Department of Nutrition and Food Sciences, University of Vermont, Burlington 05405-0044
Corresponding author: Paul Kindstedt; e-mail: paul.kindstedt{at}uvm.edu.
 |
ABSTRACT
|
|---|
Calcium lactate crystals that form white specks or haze on the surface of cheese constitute a significant quality problem for producers of Cheddar cheese. Subjective methods to evaluate crystal coverage of cheese surfaces have been reported previously, but objective methods are currently lacking. The objectives of this work were to develop and evaluate an objective method to measure the area occupied by calcium lactate crystals on surfaces of naturally smoked Cheddar cheese samples using digital photography and image analysis. Coefficients of variation ranged from 1.29 to 4.68% for 5 replicate analyses of 3 different cheese surfaces that ranged from ~2 to 49% of total surface area occupied by crystals. Thus, results showed a high degree of repeatability for the 3 cheese surfaces, which ranged from very slight and geometrically simple to very heavy and geometrically complex crystal coverage. The method underestimated total area occupied by crystals on the 3 surfaces by 0.24 to 4.83% unless the fainter crystal regions that went undetected during initial thresholding were manually segmented and quantified. The wet weight of crystal substance collected per unit of surface area from 20 different cheese samples increased exponentially as the percentage of total surface area occupied by crystals increased. These data were consistent with subjective observations that crystal regions appeared to grow vertically as well as horizontally as they expanded to occupy greater surface area. Image analysis was well suited for evaluating changes in crystal coverage during cheese aging because measurements were made nondestructively and with minimal disruption to the cheese. The area occupied by crystals on 6 different surfaces from 3 different cheese samples increased linearly (R2 = 0.94 to 0.99) during storage at 4°C for up to 33 wk. However, the rates of increase differed significantly among the 3 cheese samples. Image analysis may serve as a useful tool to quantitatively evaluate the effects of factors such as cheese composition, packaging conditions and storage temperature on rate of crystal growth and time of crystal appearance during storage.
Key Words: cheese • calcium lactate • crystal
|
INTRODUCTION
|
|---|
It is well established that calcium lactate crystals are responsible for white crystalline specks and haze that sometimes occur on the surface of Cheddar cheese (Tuckey et al., 1938; McDowall and McDowell, 1939; Shock et al., 1948; Farrer and Hollberg, 1960). These crystals are believed to form when the serum phase of the cheese becomes supersaturated with calcium and lactate ions, followed by the nucleation of calcium lactate crystals that develop into larger aggregates (Dybing et al., 1988; Kubantseva et al., 2004; Swearingen et al., 2004). They are considered visual quality defects because consumers mistake them for spoilage microorganisms. Although not harmful, these crystals are a challenge to manufacturers with regard to product reputation and financial loss due to rejected cheese (Chou et al., 2003; Swearingen et al., 2004).
Several researchers have investigated various factors associated with calcium lactate crystallization on the surface of Cheddar cheese, including high lactose levels in milk (Pearce et al., 1973); milk concentrated by ultrafiltration without diafiltration (Sutherland and Jameson, 1981); selection of certain starter culture strains (Swearingen et al., 2004); biofilm formation and contamination of cheese by nonstarter lactic acid bacteria that are able to racemize L(+) lactate to D(–) lactate (Johnson et al., 1990b; Chou et al., 2003); low storage temperature and loose packaging (Johnson et al., 1990a); low salt-in-moisture levels and low cheese pH during early aging (Swearingen et al., 2004; Rajbhandari and Kindstedt, 2005); and dehydration at the cheese surface caused by natural smoking (Rajbhandari and Kindstedt, 2005).
Although much progress has been made toward understanding some of the causes of calcium lactate crystallization and methods of prevention, cheese industry researchers (Swearingen et al., 2004) and anecdotal reports from commercial cheese manufacturers indicate that problems continue to persist in the industry. Furthermore, the basic mechanism of crystallization at the cheese surface is not completely understood. Kubantseva et al. (2004) recently characterized the solubility of calcium lactate in simple aqueous solutions, which the authors stated would provide a basis for future kinetic studies of crystallization in aqueous solutions. However, the kinetics of crystal formation and growth at the surface of cheese are likely to be more complex than in simple aqueous solutions because the protein matrix of the cheese and occluded fat globules therein impede the diffusion of dissolved substances through the serum phase (Geurts et al., 1974, 1980). Furthermore, the serum phase of cheese is complex and variable with respect to dissolved solids content (Morris et al., 1988; Swearingen et al., 2004) and may contain nucleation sites (e.g., dead bacteria) that are not present in simple aqueous solutions (Kalab, 1980). Finally, the calcium content in cheese is not completely soluble but instead exists in a pH-dependent equilibrium between the casein-associated and serum-soluble states (Kindstedt et al., 2001; Ge et al., 2002; Swearingen et al., 2004). All of these factors may affect the nucleation and growth of crystals in ways that may be difficult to predict from simple aqueous model systems.
Progress toward elucidating the mechanism of crystal formation and growth in cheese has been hindered by limitations in existing methods to quantify crystal growth. Dybing et al. (1986) described a simple, nondestructive visual method to rate the extent of crystal development (0 = no crystal development, 4 = very heavy to encrusted crystal development) on the surface of colored Cheddar cheese. This approach provided an estimate of crystal coverage, which the authors subsequently used to evaluate changes in crystal coverage during cheese aging (Dybing et al., 1988). Applying a similar subjective approach, Johnson et al. (1990b) used a 10-point scale to visually evaluate the extent of crystal development on Cheddar cheese during aging. However, little progress has been reported toward measuring crystal coverage objectively, and there continues to be a need for more accurate, repeatable, and sensitive methods to quantitatively and nondestructively measure the extent of crystal coverage on cheese surfaces.
Digital imaging technology may offer a means for quantitative measurement of surface crystals. This approach has been used widely to evaluate visual attributes of various objects, including many food applications. Image analysis has been applied successfully in cheese research to measure the size and distribution of blisters on pizza (Yun et al., 1994); correlate the size and shape characteristics of cheese shreds with other functional characteristics (Apostolopoulos and Marshall, 1994); measure topping percentage and distribution on pizza (Sun, 2000); measure the melting, browning, and oiling-off properties of Cheddar and Mozzarella cheeses as a function of time, temperature, and sample dimensions; correlate meltability with empirical test results (Wang and Sun, 2001, 2003, 2004); measure the surface area occupied by gas holes in Emmental and Ragusano cheeses (Caccamo et al., 2004); and quantify slits in Cheddar cheese (Caccamo et al., 2004). Image analysis measurements have also been shown to be more sensitive and able to differentiate cheese meltabilities more precisely than the traditional Schreiber and Arnott tests (Wang and Sun, 2002). The objectives of the present study were to develop and evaluate an image analysis method to quantify calcium lactate crystal coverage on the surface of naturally smoked Cheddar cheese. Naturally smoked cheeses are particularly well suited for image analysis because smoking renders the cheese surface orange-brown in color, which contrasts sharply with white calcium lactate crystals that may be present at the surface.
|
MATERIALS AND METHODS
|
|---|
Random weight (~300 g) retail samples of naturally smoked Cheddar cheeses exhibiting white surface crystals were obtained from local supermarkets. All cheese samples were produced at the same cheese plant, and their dates of manufacture were determined from information supplied by the manufacturer. The dimensions of a typical cheese sample were approximately 47 x 60 x 100 mm. Samples were stored at 4°C until analysis. Before the samples were photographed for image analysis, all labels were removed from the packaging film and any remaining traces of label or adhesive were gently etched away to render the film completely transparent without altering the appearance of the crystals beneath the film. After the samples were photographed, crystals from approximately one-third of the samples were carefully dislodged from the photographed surfaces using a sharp blade (1991 Trimming knife blade, Stanley Tools, New Britain, CT) and analyzed for D(–) and L(+) lactate contents by the enzymatic method described in the test kit (Boehringer Mannheim/R-Biopharm AG, test kit no. 11112821035, Darmstadt, Germany).
Photography and Image Analysis
Image generation.
A digital still camera (Sony model no. MVC-FD87, 1.3 megapixel, Sony Corp., Tokyo, Japan) was used to photograph cheese surfaces. The settings of the camera included background lights on, indoor white balance, flash off, and a resolution of 1280 x 960 pixels. The camera was mounted on a copy stand (Testrite TECS32344, Newark, NJ), with lighting provided by four 75-W lamps mounted 35 cm above the base, about 20 cm apart on opposite sides of the copy stand. Cheese samples were removed, one at a time, from refrigerated storage immediately before photographing and placed on a black cardboard base to enhance contrast. Each sample was positioned at a fixed distance (22 cm) from the camera and a steel ruler (35 cm, graduated in mm) was used to calibrate the image area. Photographs of the largest surfaces (~100 x 60 mm) were taken with the packaging film intact and saved as JPEG files. Care was taken to eliminate glare from light reflecting off the packaging film by adjusting the angles of the 4 lamps as necessary.
Image processing.
The JPEG files containing cheese images were processed using the Metamorph Offline program (Version 6.1.3, Universal Imaging Corporation, Downington, PA). Each image file was calibrated using centimeters as the unit, and a boundary line was drawn using a freehand region tool to delineate the cheese surface from the black cardboard base that constituted the rest of the image. This bounded region, which included the entire cheese surface, represented the active region where the Metamorph program functions were used to measure the area occupied by crystals. The active region was subjected to color thresholding to identify areas of crystal coverage using the Hue, Saturation, and Intensity (HSI) channel, which was found in preliminary evaluations to give better results than the 2 other channel options [i.e., Red, green, blue (RGB) or Hue, Saturation, and Luminance (HSL)] in the Metamorph program. The thresholded image was compared visually to the original image using an approach similar to that described by Caccamo et al. (2004) to confirm that thresholded regions corresponded accurately to regions occupied by crystals. In the event that glare could not be eliminated from the cheese surface during photographing, boundary lines were drawn around the affected areas to mask them using the "Unsharp Mask" function of the Metamorph program. Masking prevented regions obscured by glare from being thresholded and measured, but did not affect the measurement of crystal regions. Examples of the original and thresholded images of three cheese surfaces that exhibited different levels of crystal coverage are shown in Figure 1
. The thresholded image was then measured by image analysis and the results were expressed as the percentage of total surface area occupied by crystals.
View larger version (90K):
[in this window]
[in a new window]
|
Figure 1. Digital images of calcium lactate crystals on one surface from each of 3 samples of smoked Cheddar cheese (A, B, and C) that contained 3 levels of crystal growth. Crystals appear white on the original images (left), and green after the images were color-thresholded (right) to quantify crystal coverage by image analysis.
|
|
Method Evaluation
Repeatability.
Repeatability was evaluated by measuring 1 surface on each of 3 different cheese samples (A, B, and C) that contained different levels of crystal coverage. The 3 cheese samples were prepared and photographed as described. The original and thresholded images of the 3 cheese surfaces are shown in Figure 1
. The percentage of total surface area occupied by crystals was measured 5 times for each of the 3 images, and the means, standard deviations, and coefficients of variation were calculated. The data were analyzed by 1-way ANOVA (SuperAnova software, Abacus Concepts, Inc., Berkeley, CA) to evaluate differences in crystal coverage among the 3 images.
Thresholding error.
Because crystals do not always grow uniformly at the surface of cheese, some areas within a cheese image may have extensive crystal coverage and appear much whiter than other areas. Typically, thresholding readily distinguishes whiter crystallized regions, but may fail to distinguish areas with faint crystal coverage, resulting in an underestimation of the total area occupied by crystals. The following approach was used to evaluate the extent of underestimation due to thresholding error for the 3 cheese surfaces shown in Figure 1. First, the active region of each image was thresholded (see Figure 1), and the thresholded areas (representing the regions that were identified as crystals) were measured. The thresholded images were then segmented; that is, regions occupied by faint crystals that were not thresholded (therefore not measured) were identified by visual inspection, boundary lines were drawn around the regions of interest using a freehand region tool of the program, and the bounded regions were thresholded and measured. The total area occupied by the segmented crystals was determined. Each of the 3 images was analyzed 5 times for thresholded crystal area and total segmented crystal area.
Thresholding error was defined as follows:
Method Application
Relationship between crystal wet weight and area.
Twenty-nine surfaces from 20 different cheese samples were photographed and the images were analyzed as described to determine the percentage of total area occupied by crystals. The photographed samples were returned to refrigerated storage and briefly tempered to 4°C, after which the packaging film was removed from the cheese. The crystals were then carefully dislodged from the photographed surfaces using a sharp blade (1991 Trimming knife blade) and weighed. The relationship between crystal wet weight, expressed per unit area of cheese surface, and crystal coverage, expressed as the percentage of total cheese surface area occupied by crystals, was evaluated using linear and nonlinear regression analyses (SuperAnova software, Abacus Concepts, Inc.).
Aging study.
Three samples of smoked Cheddar cheese, ranging in age from 20 to 25 wk after manufacture, were obtained from retail sources. Two of the samples (A and B) originated from the same vat of cheese and exhibited very slight or no crystal formation at the cheese surfaces. The third sample (C) was produced on a different day and displayed pronounced crystal coverage on its surfaces. The samples were stored at 4°C and photographs of the 2 largest (opposite) surfaces (~100 x 60 mm) of each cheese were taken periodically during storage. The photographs were taken with the packaging film intact. It took less than 1 min to photograph each cheese sample, after which the samples were immediately returned to 4°C storage until the next photography session. Samples A and B were photographed at 25, 28, 31, 34, 37, 40, 43, 46, 49, 52, 55, and 58 wk after manufacture. Sample C was photographed at 20, 23, 26, 29, 32, and 35 wk after manufacture. Changes in crystal coverage over time for each of the 2 surfaces for each cheese were evaluated using linear regression analysis (SuperAnova software, Abacus Concepts, Inc.)
|
RESULTS AND DISCUSSION
|
|---|
Image analysis works best when there is a sharp contrast between the objects of interest and the background. Naturally smoked cheese samples provided excellent contrast for the measurement of surface crystals (i.e., white crystals against an orange-brown surface), which simplified the image analysis procedure by not having to adjust the contrast of the original image. However, it is worth noting that preliminary studies indicated that the procedure described for naturally smoked cheese also appears to hold promise for use with uncolored (white) cheese, provided that the contrast between the light yellow cheese background and the white crystals is enhanced using the "Adjust Digital Contrast" function of the Metamorph program (data not shown). Additional research is needed to fully evaluate the potential of the image analysis method for this application.
The 2 largest (~100 x 60 mm) opposite surfaces of the retail samples were chosen for evaluation in this study because the smaller side surfaces often contained creases and seams in the packaging film that could not be differentiated from the surface crystals. In contrast, the packaging film on the large surfaces was smooth and transparent. An important feature of image analysis is that cheese samples can be easily and quickly photographed with the packaging film intact and then returned to controlled storage conditions. Therefore, measurements can be made nondestructively and with minimal disruption to the cheese, making image analysis well suited for evaluating changes in crystal coverage over time in retail cuts of cheese, as in this study, or potentially in large blocks of cheese that are wrapped in clear film packaging.
The analyses of crystals collected from about one-third of the samples that were included in this work indicated that the crystals contained from 18 to 52% L(+) lactate. No D(–) lactate was detected. Thus, the data support the view that calcium lactate was the major component of the white crystalline substance present on the cheese surfaces. In a different study of crystallization on retail cuts of smoked Cheddar cheeses that were produced at the same cheese plant as those in the present study, the authors reported that the white surface crystals varied considerably in composition, with mean values of 8.1% Ca, 52.1% total lactate, 28.5% water, 0.17% P, 1.56% NaCl, and 8.9% CP (Rajbhandari and Kindstedt, 2005). Factors that cause crystal composition to vary, and the effect of compositional variation on crystal structure and appearance, have not been investigated.
Method Evaluation
Repeatability.
The results from 5 replicate analyses of crystal coverage on the 3 different cheese surfaces shown in Figure 1 are presented in Table 1. Crystal coverage differed significantly among the 3 surfaces. The mean percentage of total surface area occupied by crystals was 1.86, 16.64, and 49.25% for surfaces A, B, and C, respectively. Coefficients of variation were less than 5% for all 3 surfaces, ranging from 1.29% (C) to 4.68% (B). Thus, the analyses showed a high degree of repeatability for surfaces that ranged from very slight and geometrically simple (A) to very heavy and geometrically complex (C) crystal coverage. It should be noted, however, that the distribution of crystals on cheese surfaces is not always uniform. For example, most of the crystals on surface B were concentrated in one region; therefore, it was necessary to analyze the entire surface to obtain a reliable measure of crystal coverage on a specific surface. Furthermore, it is likely that different surfaces of the same block of cheese may develop different levels of crystal coverage over time due to factors such as differences in the tightness of the packaging film or temperature history among the surfaces.
View this table:
[in this window]
[in a new window]
|
Table 1. Repeatability (mean, SD, coefficient of variation) of image analysis estimates of the percentage coverage of calcium lactate crystals on one surface from each of 3 samples of smoked Cheddar cheese that contained 3 levels of crystal growth: (A) about 2.0%, (B) about 16%, and (C) about 49% crystal coverage.
|
|
Thresholding error.
Thresholding error refers to the underestimation of crystal coverage that occurs when areas occupied by faint crystals are not thresholded and therefore not measured. Factors that cause some crystallized regions to be fainter (i.e., less white) than others have not been addressed in the literature but may involve the thickness (depth) and density of the crystal mass, as discussed below. This unmeasured area of faint crystals can be quantified after the initial thresholding by segmenting the area as described earlier. The thresholding errors obtained during 5 replicate analyses of the 3 images shown in Figure 1 are presented in Table 2. The mean thresholding error ranged from 0.24% (C) to 4.83% (B). Thus, the extent of underestimation due to thresholding error was small relative to the area being measured. It is evident that crystal coverage is most accurately measured by image analysis when the thresholded plus the segmented areas are combined; however, thresholding these segmented areas can be tedious and time-consuming, making the method less user friendly. Alternatively, for many applications, the thresholding error may be considered small enough to neglect, obviating the need for segmentation. The data presented below were obtained from images that were thresholded but not segmented, meaning that the results did not account for thresholding error.
View this table:
[in this window]
[in a new window]
|
Table 2. Underestimation of the area of crystals on one surface from each of 3 samples of smoked Cheddar cheese (A, B, and C) due to incomplete thresholding of fainter crystal regions.
|
|
Method Application
Relationship between crystal wet weight and area.
The data presented in Figure 2 indicate that the wet weight of crystals obtained per unit area of cheese surface generally increased in a nonlinear (i.e., exponential; R2 = 0.63) manner as the percentage of total surface area occupied by crystals increased. This pattern was unexpected. The authors anticipated that the force of the packaging film on cheese surface would largely prevent vertical growth. Therefore, it was assumed that crystals would grow horizontally to a uniform depth, resulting in a linear relationship between crystal wet weight per unit area and area of coverage. However, subjective observations of crystallized regions as they changed over time suggested that crystals tended to grow vertically as well as horizontally, thereby causing crystallized regions to attain greater thickness as the regions expanded horizontally and occupied more of the total surface area. However, the thickness of crystallized regions also appeared to vary considerably among cheese samples that had extensive crystal coverage, which may have contributed to the large variation in the data shown in Figure 2. There appeared to be sample-to-sample differences in the density and structure of crystallized regions. Some heavily crystallized samples contained white regions that were subjectively characterized as having a light and open structure, whereas white regions on other heavily crystallized samples appeared more dense and compact. Factors that govern the structure and density of surface crystals and their growth horizontally and vertically are not well understood. Further research is needed to elucidate crystal growth behavior and to determine whether differences in thickness and density affect the whiteness of crystallized regions. The proposed image analysis method measures only crystal area and does not take into account potential differences in crystal thickness and density. Different analytical approaches will be needed to address these factors.
View larger version (11K):
[in this window]
[in a new window]
|
Figure 2. Relationship between crystal coverage (percentage of total surface area occupied by crystals) and the wet weight of crystal substance collected from the surface (per unit of surface area). Data represent 29 surfaces from 20 different samples of naturally smoked Cheddar cheese.
|
|
Aging study.
Changes in the percentage of total surface area occupied by crystals on 3 cheese samples (A, B, and C) during 4°C storage are shown in Figure 3. Crystal coverage was measured on 2 opposite surfaces (1 and 2) for each cheese sample. For all 6 surfaces, crystal coverage increased in a linear manner, with R2 values ranging from 0.94 to 0.99. For each of the 3 cheese samples, the slopes of the regression equations for the 2 opposite surfaces did not differ significantly. Thus, crystal growth, as measured by increases in surface area, did not differ significantly among different surfaces of the same cheese sample. However, there were significant differences in the slopes of the linear regression equations between different cheese samples. Specifically, the regression slopes for both surfaces of sample C (C1 and C2) were significantly greater than those of sample A (A1 and A2) or sample B (B1 and B2). In addition, the regression slope for surface 2 of sample B (B2) was significantly greater than that of surface 1 of sample A (A1). Thus, crystal growth rates differed among the 3 different cheese samples. Factors that influence the rate of crystal growth on cheese surfaces are not completely understood. Presumably, the linear increases in crystal coverage shown in Figure 3 would have eventually reached maximum values and leveled off, had the study been extended. It is not known whether all 6 surfaces would have attained the same maximum level of coverage.
Estimates of the time at which crystals first became visible on the cheese surfaces were obtained by calculating the y-intercepts of the regression lines, represented by the dashed lines shown in Figure 3. Based on these extrapolations, crystals became visible on the surfaces of samples A and B between wk 20 and 25 after manufacture, whereas the estimated onset of visible crystals occurred much earlier (~5 wk after manufacture) for sample C. According to the manufacturer, the cheese samples underwent natural smoking after 2 to 4 mo of aging. Therefore, crystals probably first appeared on the surfaces of samples A and B several weeks after the smoking treatment, whereas crystals in sample C may have appeared before the cheese was subjected to natural smoking. Other researchers also reported differences in the time that crystals first appeared on different cheese samples (Johnson et al., 1990a,b; Chou et al., 2003). The factors that influence the time at which crystals first appear on the surface of cheese, and the relationship between the time of first crystal appearance and their subsequent rate of growth are not well understood. Clearly, early onset of visible crystals combined with rapid subsequent growth, such as occurred in sample C (Figure 3), is particularly detrimental to product quality and shelf life.
|
CONCLUSIONS
|
|---|
A method consisting of digital photography combined with image analysis was used to quantify the area occupied by calcium lactate crystals on the surface of naturally smoked Cheddar cheese samples. The analyses showed a high degree of repeatability for surfaces that ranged from very slight and geometrically simple to very heavy and geometrically complex crystal coverage. The method slightly underestimated the total area occupied by crystals unless manual segmentation was used to quantify faint crystal regions that went undetected during the initial thresholding of crystals. The wet weight of crystal substance collected per unit of cheese surface area increased exponentially as the percentage of total surface area occupied by crystals increased. This was consistent with subjective observations that crystal regions appeared to grow vertically as they expanded horizontally to occupy greater surface area. Area occupied by crystals expanded in a linear manner but at different rates for cheese samples stored at 4°C with their packaging film intact. Image analysis may serve as a useful tool for quantitatively evaluating the effects of factors such as cheese composition, packaging conditions, and storage temperature on rate of crystal growth and time of crystal appearance during storage.
|
ACKNOWLEDGEMENTS
|
|---|
The financial support of the Northeast Dairy Foods Research Center and USDA Hatch Project VT-H01010 are gratefully acknowledged.
|
FOOTNOTES
|
|---|
* Use of names and identification of specific models of equipment and software is for scientific clarity and does not constitute any endorsement of product by the authors, the University of Vermont, or the Northeast Dairy Foods Research Center. 
Received for publication June 1, 2005.
Accepted for publication August 4, 2005.
|
REFERENCES
|
|---|
Apostolopoulos, C., and R. J. Marshall. 1994. A quantitative method for the determination of shredability of cheese. J. Food Qual. 17:115–128.
Caccamo, M., C. Melilli, D. M. Barbano, G. Portelli, G. Marino, and G. Licitra. 2004. Measurement of gas holes and mechanical openness in cheese by image analysis. J. Dairy Sci. 87:739–748.[Abstract/Free Full Text]
Chou, Y.-E., C. G. Edwards, L. O. Luedecke, M. P. Bates, and S. Clark. 2003. Nonstarter lactic acid bacteria and aging temperature affect calcium lactate crystallization in Cheddar cheese. J. Dairy Sci. 86:2516–2524.[Abstract/Free Full Text]
Dybing, S. T., S. A. Brudvig, J. A. Wiegand, and E. A. Huang. 1986. A simple method for estimating the extent of surface crystal development on colored Cheddar Cheese. J. Food Prot. 49:421–422.
Dybing, S. T., J. A. Wiegand, S. A. Brudvig, E. A. Huang, and R. C. Chandan. 1988. Effect of processing variables on the formation of calcium lactate crystals on Cheddar cheese. J. Dairy Sci. 71:1701–1710.[Abstract/Free Full Text]
Farrer, K. T. H., and W. C. J. Hollberg. 1960. Calcium lactate on rindless cheese. Aust. J. Dairy Technol. 15:151–152.
Ge, Q., M. Almena-Aliste, and P. S. Kindstedt. 2002. Reversibility of pH-induced changes in the calcium distribution and melting characteristics of Mozzarella cheese. Aust. J. Dairy Technol. 57:3–9.
Geurts, T. J., P. Walstra, and H. Mulder. 1974. Transport of salt and water during salting of cheese. 1. Analysis of processes involved. Neth. Milk Dairy J. 28:102–129.
Geurts, T. J., P. Walstra, and H. Mulder. 1980. Transport of salt and water during salting of cheese. 2. Quantities of salt taken up and of moisture lost. Neth. Milk Dairy J. 34:229–254.
Johnson, M. E., B. A. Riesterer, and N. F. Olson. 1990a. Influence of nonstarter bacteria on calcium lactate crystallization on the surface of Cheddar cheese. J. Dairy Sci. 73:1145–1149.[Abstract/Free Full Text]
Johnson, M. E., B. A. Riesterer, C. Chen, B. Tricomi, and N. F. Olson. 1990b. Effect of packaging and storage conditions on calcium lactate crystallization on the surface of Cheddar cheese. J. Dairy Sci. 73:3033–3041.[Abstract]
Kalab, M. 1980. Decayed lactic bacteria—A possible source of crystallization nuclei in cheese. J. Dairy Sci. 63:301–304.[Abstract/Free Full Text]
Kindstedt, P. S., A. Zielinski, M. Almena-Aliste, and C. Ge. 2001. A post-manufacture method to evaluate the effect of pH on Mozzarella cheese characteristics. Aust. J. Dairy Technol. 56:202–207.
Kubantseva, N., R. W. Hartel, and P. A. Swearingen. 2004. Factors affecting solubility of calcium lactate in aqueous solutions. J. Dairy Sci. 87:863–867.[Abstract/Free Full Text]
McDowall, F. H., and A. K. R. McDowell. 1939. The white particles in mature Cheddar cheese. J. Dairy Res. 10:118–119.
Morris, H. A., C. Holt, B. E. Brooker, J. M. Banks, and W. Manson. 1988. Inorganic constituents of cheese: Analysis of juice from a one-month-old Cheddar cheese and the use of light and electron microscopy to characterize the crystalline phases. J. Dairy Res. 55:255–268.
Pearce, K. N., L. K. Creamer, and J. Gilles. 1973. Calcium lactate deposits on rindless Cheddar cheese. N.Z. J. Dairy Sci. Technol. 8:3–7.
Rajbhandari, P., and P. S. Kindstedt. 2005. Compositional factors associated with calcium lactate crystallization in smoked Cheddar cheese. J. Dairy Sci. 88:3737–3744.[Abstract/Free Full Text]
Shock, A. A., W. J. Harper, A. M. Swanson, and H. H. Sommer. 1948. What’s in those "white specks" on Cheddar? Wisc Agr. Exp. Sta. Bull. 474. Univ. Wisconsin, Madison.
Sun, D.-W. 2000. Inspecting pizza topping percentage and distribution by a computer vision method. J. Food Eng. 44:245–249.
Sutherland, B. J., and G. W. Jameson. 1981. Composition of hard cheese manufactured by ultrafiltration. Aust. J. Dairy Technol. 36:136–142.
Swearingen, P. A., D. E. Adams, and T. L. Lensmire. 2004. Factors affecting calcium lactate and liquid expulsion defects in Cheddar cheese. J. Dairy Sci. 87:574–582.[Abstract/Free Full Text]
Tuckey, S. L., H. A. Ruehe, and G. L. Clark. 1938. X-ray diffraction analysis of white specks in Cheddar cheese. J. Dairy Sci. Abs. 21:161.
Wang, H. H., and D.-W. Sun. 2001. Evaluation of the functional properties of Cheddar cheese using a computer vision method. J. Food Eng. 49:49–53.
Wang, H. H., and D.-W. Sun. 2002. Correlation between cheese melt-ability determined with a computer vision method and with Arnott and Schreiber tests. J. Food Sci. 67:745–749.
Wang, H. H., and D.-W. Sun. 2003. Assessment of cheese browning affected by baking conditions using computer vision. J. Food Eng. 56:339–345.
Wang, H. H., and D.-W. Sun. 2004. Evaluation of the oiling off property of cheese with computer vision: Influence of cooking conditions and sample dimensions. J. Food Eng. 61:57–66.
Yun, J. J., D. M. Barbano, E. F. Bond, and M. Kalab. 1994. Image analysis method to measure blister size and distribution on pizza. Pages 33–38 in Proc. 31st Annu. Marschall Italian Speciality Cheese Seminar, Madison, WI. Rhône-Poulenc, Madison, WI.
This article has been cited by other articles:
|
|
|
|
 
P. Rajbhandari, J. Patel, E. Valentine, and P. S. Kindstedt
Chemical changes that predispose smoked Cheddar cheese to calcium lactate crystallization
J Dairy Sci,
August 1, 2009;
92(8):
3616 - 3622.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
P. Rajbhandari and P. S. Kindstedt
Characterization of Calcium Lactate Crystals on Cheddar Cheese by Image Analysis
J Dairy Sci,
June 1, 2008;
91(6):
2190 - 2195.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
); B1 (
); B2 (
); C1 (
), C2 (
). Linear regression lines are shown for the 6 surfaces. The slopes of the regression equations did not differ significantly between surfaces on the same cheese sample. However, the slopes of the regression equations for sample C (C1 and C2) were significantly greater than those from sample A (A1 and A2) or sample B (B1 and B2). Also, the slope of the regression equation for surface B2 was significantly greater than that for surface A1. Dashed lines indicate the estimated times that the crystals first became visible, based on the y-intercepts calculated from the regression equations.