Species Wise Classification of Crust Leather images based on histogram equalization

The classification of Crust leather on various animal species is one of the sorting processes in leather production. At the Industry under the sorting process, segregation of Crust leather on the basis of the type of leather made on Hides from Cattle, Buffalo and Skins from, Goat and Sheep. This process is normally does by a team of experts through manually and it consumes good amount of time. Hence, authors have proposed a new dataset namely INT_LEA_IMG_CLRI_UNOM with 609 images consisting of Training and Testing dataset. The images are pre-processed using filters such as LAPLACIAN, MEDIAN, SOBEL, ROBERTS, SCHARR, PREWITT, GAUSSIAN and the image enhanced by employing of Contrast Limited Adaptive Histogram Equalization (CLAHE). Features are extracted using Structural Similarity Index (SSIM) and Histogram. Further features are selected through Random Forest and Support Vector Machines. Classification accuracy is confirmed at significant level of the proposed method using Naïve Bayes classifier at 94 % level of accuracy. Hence current study confirming this method will be complement to the sorting process while segregation of Crust leather on the basis of the type of leather made on Hides from Cattle, Buffalo and Skins from, Goat and Sheep.