Extreme Inception Model and Its Application in Indian Food Image Classification

In our increasingly digital world, the ability to analyze and interpret images has become indispensable. Image classification, a fundamental task in computer vision, plays a pivotal role in various domains including healthcare, automotive, security, and entertainment. Transfer learning, particularly useful when data is limited, has become a popular approach for optimizing pre-trained models for specific tasks. One prominent model in this area is Xception (Extreme Inception), which represents a significant advancement in deep learning architectures. Xception provides a powerful alternative to traditional convolutional neural networks (CNNs) for image-related tasks, delivering strong performance while being computationally efficient. An important factor in situations where high precision is needed when computational resources are scarce is Xception’s robust performance and economical architecture. The use of the Xception model with ImageNet weight is indicated in this work. Here, the database of the images comprising of 18 distinct Indian food images for training and validation purpose. The images are preprocess using data augmentation process the proposed model Yields 99 percent accuracy which is highly acceptable. Additionally, the achieved accuracy is evaluated against the state-of-the-art methods, showing that our exception model outperforms all other techniques.