Search

Contrast enhancement algorithms have been evolved through last decades to meet the requirement of its objectives. Actually, there are two main objectives while enhancing the contrast of an image: (i) improve its appearance for visual interpretation and (ii) facilitate/increase the performance of subsequent tasks (e.g., image analysis, object detection, and image segmentation). Most of the contrast enhancement techniques are based on histogram modifications, which can be performed globally or locally. The Contrast Limited Adaptive Histogram Equalization (CLAHE) is a method which can overcome the limitations of global approaches by performing local contrast enhancement. However, this method relies on two essential hyperparameters: the number of tiles and the clip limit. An improper hyperparameter selection may heavily decrease the image quality toward its degradation. Considering the lack of methods to efficiently determine these hyperparameters, this article presents a learning-based hyperparameter selection method for the CLAHE technique. The proposed supervised method was built and evaluated using contrast distortions from well-known image quality assessment datasets. Also, we introduce a more challenging dataset containing over 6200 images with a large range of contrast and intensity variations. The results show the efficiency of the proposed approach in predicting CLAHE hyperparameters with up to 0.014 RMSE and 0.935 R2 values. Also, our method overcomes both experimented baselines by enhancing image contrast while keeping its natural aspect. [ABSTRACT FROM AUTHOR]

• We identify different grades of fermentation of cocoa beans using computer vision. • Image features were useful for classification of cocoa beans. • Random decision forests demonstrated high accuracy and precision for classification models. • Computer vision may be an analytical method to identify grades of fermentation. • The proposed method has potential application in the cocoa and chocolate industry. Fermentation of cocoa beans is a critical step for chocolate manufacturing, since fermentation influences the development of flavour, affecting components such as free amino acids, peptides and sugars. The degree of fermentation is determined by visual inspection of changes in the internal colour and texture of beans, through the cut-test. Although considered standard for evaluation of fermentation in cocoa beans, this method is time consuming and relies on specialized personnel. Therefore, this study aims to classify fermented cocoa beans using computer vision as a fast and accurate method. Imaging and image analysis provides hand-crafted features computed from the beans, that were used as predictors in random decision forests to classify the samples. A total of 1800 beans were classified into four grades of fermentation. Concerning all image features, 0.93 of accuracy was obtained for validation of unbalanced dataset, with precision of 0.85, recall of 0.81. Although the unbalanced dataset represents actual variation of fermentation, the method was tested for a balanced dataset, to investigate the influence of a smaller number of samples per class, obtaining 0.92, 0.92 and 0.90 for accuracy, precision and recall, respectively. The technique can evolve into an industrial application with a proper integration framework, substituting the traditional method to classify fermented cocoa beans. [ABSTRACT FROM AUTHOR]

Veins in pork thigh carcass are directly related to the quality of dry-cured ham, and consequently to its market value. Some veining defects over the surface of raw ham are easily detected by humans and precisely assessed by a specialist. However, the automatic evaluation of raw ham quality by image analysis poses some challenges to the traditional Computer Vision Systems (CVS), many of them grounded on the complex image pattern related to each defect level. To improve the CVS classification performance without overburdening feature extraction, as well as the common machine learning modelling, we propose Dual Stage Image Analysis (DSIA). DSIA is an additional step in a CVS, that was built in two stages based on the "divide and conquer" strategy. The first stage consists of splitting the region of interest into sub-regions to predict the presence of veining. In the second stage, the algorithm computes the number of veining sub-regions to assess the final defect level classification. A total of 194 raw ham samples were used to evaluate the DSIA performance in the experiments. Support Vector Machine and Random Forest algorithms were compared for inducing the classification model using 92 image features. Random Forest model was the best, capable of predicting defect level with 88.10% accuracy using DSIA. Without DSIA, the CVS with RF achieved an accuracy of 63.10%. • The proposed method improved the image pattern recognition of complex patterns. • The classification accuracy was improved by more than 20%. • A real dataset was explored with two different machine learning algorithms. • It is a straightforward method to perform embedded in the traditional CVS pipeline. [ABSTRACT FROM AUTHOR]