Acadlore takes over the publication of IJCMEM from 2025 Vol. 13, No. 3. The preceding volumes were published under a CC BY 4.0 license by the previous owner, and displayed here as agreed between Acadlore and the previous owner. ✯ : This issue/volume is not published by Acadlore.
Open AccessIndianFoodNet: Detecting Indian Food Items Using Deep Learning
View Full Article|
Download PDFAbstract:
India is widely recognized for its wealthy heritage, subculture and myriad Indian cuisines. Indian Cuisines are famous around the globe for their taste and flavors. Indian Cuisines detection using computer vision-based methods has been limited till now because of the absence of a standard dataset needed to inspect the deep learning-based object detection models for detecting Indian Food Cuisine using electronic devices. Measuring food quantities in each item are very challenging tasks for a person. In this study the dataset IndianFoodNet has been introduced, containing more than 5500 high-quality images and 5000+ annotations spreading across thirty classes of Indian food items. A comparative study of various state-of-the-art object detection models- YOLO5, YOLO7 and YOLO8 has been provided in the study. Further, the model performance has been inspected and evaluated (As in training summary of YOLO at 5 epochs YOLO8 precision is 0.775 higher than precision of YOLO7 and YOLO5.Recall value of YOLO7 is least in comparison with YOLO5 having value 0.671 and YOLO8 having recall value 0.719) by qualitatively analyzing the prognostic made on the images of the dataset which are segregate for testing.
