AUTOMATED FOOD IMAGE CLASSIFICATION USING DEEP LEARNING APPROACH

Computer Engineering

Deep learning-based automated food image classification has various purposes, including calorie estimate, diet monitoring, and food safety inspection. This field of work is now under rapid development. On challenges involving the categorization of food images, deep learning models have been proven to outperform conventional machine learning techniques and yield leading-edge findings. This presentation offers a quick review of the subject of automated food image categorization using deep learning algorithms. The Proposed work talks about classifying food images, the variety of deep learning techniques that have been utilised, and the most significant recent developments in the field. The wide variety of food picture is one of the primary concerns in food image categorization. Food items can be displayed in a variety of shapes, dimensions, colours, and textures, and it can be challenging for a computer to figure out between them, particularly when there is noise in the image or when the food items are partially obscured. Deep learning algorithms can overcome these obstacles by learning to extract complicated information from food photos. After that, a model is trained using these attributes to categorise food products into various categories. Several deep learning techniques have been applied to the categorization of food images. Convolutional neural networks (CNNs) are one popular method. CNNs are ideal for image classification problems because they can learn spatial information from pictures. Transfer learning is another popular strategy. Transfer learning is the process of using a deep learning model that has already been trained on a sizable picture dataset, such as ImageNet. The pre-trained model is then refined using smaller scales dataset of food pictures. Using deeper and more sophisticated CNN architectures together with data augmentation approaches to expand the quantity and variety of the training dataset have led to recent advancements in the field of food image categorization. the accuracy of the planned work using CNN – Mobilenet V2 is 90.5%, which is greater than previous related work using CNN – Imagenet is 86.6%. Altogether, deep learning approaches for automatic food picture categorization are a fast-emerging topic containing numerous potential applications

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.