Abstract This article presents the design of a robust leather species identification technique. It aims to intertwine deep learning with leather image analysis. Hence, this work collects and analyzes large-scale leather image data for diverse learning. The data involve 7600 unique images with species-distinct and varied pore patterns from four species. It proposes a novel dual-stream architecture for accurate leather image classification. It is a fusion of local binary pattern-based texture analysis and MobileNet-based adaptive feature learning, hence the name LBPMobileNet. The former highlights the local structural pattern of an image, and the latter efficiently learns the species’ uniqueness. The dual-stream model analyzes two sources of images to provide more reliable and robust learning from different textured images. At the same time, it adopts two MobileNets to design a computationally efficient model. Thus, the proposed model utilizes limited resources and provides 96.45% accurate leather image classification. Further, the performance analysis affirms the generalization ability of the proposed model by predicting species from leather images with ideal and complex behavior. It also validates the robustness and computational efficiency of the proposed model with the state-of-the-art deep learning models. Thus, this study proves the relevance of local binary patterns, fused feature analysis, dual-stream architecture, and deep learning for efficient leather image analysis. It, thereby, assists the leather experts by developing an automatic and accurate species prediction method.

This paper introduces deep learning (DL) for leather species identification. It exploits the application of transfer learning on the existing Convolutional Neural Networks (ConvNets). The application of transfer learning fine-tunes the ConvNet parameters to learn the novel leather image data. This research investigates the performance of four ConvNets, namely, AlexNet, VGG16, GoogLeNet, and ResNet18, to predict the leather species. The comparative study affirms the efficacy of ResNet18 in learning the complex pore structural behavior of leather images. It efficiently classifies the leather images into four respective species with the highest accuracy (99.69%). It outperforms the existing ML-based prediction with a 7% improvement. Therefore, ConvNet is the best solution to deal with inter-species similarity and intra-species variability, the practical challenges of the leather images. It thus develops a fully-automated leather species identification technique that paves the way for biodiversity preservation and consumer protection