FOOD CALORIE ESTIMATION USING DEEP LEARNING AND COMPUTER VISION

Volume-09 Issue 04, April-2025 ISSN: 2456-9348 Impact Factor: 8.232 International Journal of Engineering Technology Research & Management Published By: https://www.ijetrm.com/ IJETRM (http://ijetrm.com/) [1] FOOD CALORIE ESTIMATION USING DEEP LEARNING AND COMPUTER VISION Syed Khaliq Assistant Professor, dept of AI&ML, JBIET College, Hyderabad, India [email protected] Y Jayasurya Reddy, Y Sriyukth Chowdary Students, dept of AI&ML, JBIET College, Hyderabad, India [email protected] , [email protected] ABSTRACT In later a long time, progressions in counterfeit insights and computer vision have empowered computerized arrangements for wellbeing observing and dietary evaluation. In this article, a deep learning-based system for estimating food calories is presented with the help of object recognition and image processing. The proposed system uses the Yolov5 model for food detection, OpenCV for the preprocessing of the interactive user interface, and power supplies. By using pre-formed models and pre-defined calorie data sets, the system recognizes food and provides approximate calorie counts based on volume estimation techniques. This paper presents a profound learning-based framework for evaluating nourishment calories utilizing question location and picture handling. The proposed framework utilizes the YOLOv5 demonstrate for nourishment acknowledgment, OpenCV for picture preprocessing, and Stream lit for an intelligently client interface. The comes about demonstrate that this approach is productive and adaptable for real-time applications in individual wellbeing tracking. Keywords: Nourishment calorie estimation, profound learning, YOLOv5, OpenCV, Stream lit, protest discovery, computer vision, dietary examination. 1. INTRODUCTION Dietary monitoring is important for maintaining a healthy lifestyle. It is especially important for those who control weight -related conditions such as obesity and diabetes. With the recent achievements of computer vision and deep learning, we were able to automate this process by evaluating calories based on images. Body is a disease that represents a high ratio of muscle to fat. Combustion of many calories is one of these reasons. The body stores excessive calories in the ratio of muscle and fat. People need to monitor calorie consumption to increase their shape or to maintain healthy weight. But this interaction can be disappointed and tired. People often do not observe food consumption. People need to monitor calorie consumption to increase their shape or to maintain healthy weight. This request was done to simplify the following: People want to know how much of the food they spend as well as the appearance of food in the picture. Finally, based on the volume predicted by the model, we determine the content of the calories of food. In most cases, however, people have problems with the evaluation and measurement of the amount of food they eat. In this study, we use a method of analysing each design network and recognizing a picture based on deep learning to increase the accuracy of nutritional evaluation. Content of calories of food. 2. LITERATURE SURVEY In several studies, we used deep training to study food recognition and calorie evaluation. Some studies used deep training to investigate the detection of food and calorie assessments. Nevertheless, the Yolo model has a real-time reputation for its functionality. Additional research highlights the role of mobile applications and cloud computing in nutrition monitoring. While many existing applications rely on passive food registration, the performance of systems controlled by artificial intelligence aims to minimize user effort across accuracy. Imagebased systems that can identify several foods have been proposed as a solution to this problem. He has the Volume-09 Issue 04, April-2025 ISSN: 2456-9348 Impact Factor: 8.232 International Journal of Engineering Technology Research & Management Published By: https://www.ijetrm.com/ IJETRM (http://ijetrm.com/) [2] advantage of being able to recognize some photos, but it always depends on the device supported in stereo mode. A new model is then provided to create a distribution card from the food to estimate the size of each portion of the food in the entire input image. It also processes images faster. Nevertheless, only the food parts class is performed at the same time. Cage Yanai and Koichi Okamoto [1] Tokyo Electric and Architecture University of Japan Ministry of Information and Science «, the automated system uses an external recognition server to evaluate smartphone calories and not calories in images. [2] Integrated Systems for Food Analysis Heng Peng Runny Mao, Zemon Shao, Janine L. Wright, Deborah A. Kerr, Carol J. Bushi and Fing Zhu all contributed to this article. Automatic analysis of image-based performance products can be used to process or multitasking (such as recognition or partial evaluation). [3] Actual evaluation of images of each image for accurate calorie calculation for calorie assessment by images that can identify different products. 3.PROBLEM SYSTEM The proposed system includes three main components: food detection, volume grade and graphic user interface. Workflow includes: Image Receive: The user uploads the power image through the web application based on the stream time. Food detection: The YOLOV5 model identifies food from the image. Calorie Assessment: Using a predetermined density and calorie value, the system calculates the estimated calorie content. The system charges a product detected by the appropriate calorie grade. The goal is to capture the type of food. Evaluate the calories of foods. Notify users whether food consumption is evenly prescribed. C. Food detection in the proposed system is an existing idea that allows food to be recognized and recognized according to the input image. Our model was trained as Category 101 foods. Furthermore, the idea is to assess the calorie content in food detection. Remitted neural networks (CNNs) are used to recognize food. Food weights are provided by input and the exact calorie cost of the food is calculated according to standard calorie values. There is no description of the VI software. With the project, the inventor app will democratize software development, expanding all features, especially the young people, and making changes from technological consumption to technology. Python is a multipurpose programming language that can be used for modelling, creating websites, and communicating with database systems. TensorFlow is a free library with open-source software for machine learning and artificial intelligence. This is extremely important in modern systems. It can be used to process images and videos to identify objects, faces, or human writing. In various libraries such as NumPy, Python can manage OpenCV array structures for analysis. Use vector rooms to determine image templates and various features and perform mathematical operations for these features. 4. METHODOLOGY • Discovery: YOLOV5 Show is used to distinguish the food of portable photos. • Preprocessing: OpenCV is utilized to improve picture clarity and normalize input dimensions • Volume Estimation: A reference question, such as a coin or thumb, is utilized for scale estimation. • User Interface: A Stream lit application empowers clients to connected with the framework by uploading pictures and accepting moment calorie estimates. Model Preparing: The YOLOv5 show was fine-tuned on freely accessible nourishment datasets such as Food-101 and UEC-FOOD100 to make strides classification exactness. • Model preparation: YOLOV5 shows that it has been expanded as fine as Food-1010101010101111. 5. EXPERIMENTAL RESULTS The system was tested on a dataset comprising common food items, and its performance was evaluated based on detection accuracy and calorie estimation precision. The YOLOv5 model achieved an average detection accuracy of 92%. The calorie estimates were compared with nutritional databases, showing a deviation of approximately 1015%, which is acceptable for dietary assessment purposes. Further analysis revealed that lighting conditions and background uniformity significantly influenced detection accuracy. Controlled environments with sufficient lighting produced more reliable calorie estimations. The system was also tested with different reference objects to determine the most effective scale estimation approach, with a Volume-09 Issue 04, April-2025 ISSN: 2456-9348 Impact Factor: 8.232 International Journal of Engineering Technology Research & Management Published By: https://www.ijetrm.com/ IJETRM (http://ijetrm.com/) [3] standardized object (coin or thumb) yielding the best results. 6. Conclusion and Future Work This paper presents a novel approach to automated calorie estimation using deep learning and computer vision. The proposed system offers a user-friendly and efficient way to monitor dietary intake. Future work will focus on improving segmentation accuracy, incorporating additional food categories, and integrating a real-time mobile application. The future work will focus on increasing the accuracy of segmentation, including the food category and mobile application integration of additional products. 7. REFERENCES 1. Redmon, J., & Farhadi, A. (2018). YOLOv3: An Incremental Improvement. arXiv preprint arXiv:1804.02767. 2. Liu, W., Anguelov, D., Erhan, D., et al. (2016). "SSD: Single Shot MultiBox Detector." European Conference on Computer Vision (ECCV). 3. Bochkovskiy, A., Wang, C-Y., & Liao, H-Y. M. (2020). "YOLOv4: Optimal Speed and Accuracy of Object Detection." arXiv preprint arXiv:2004.10934. 4. Pouladzadeh, P., Shirmohammadi, S., & Almaghrabi, R. (2014). Measuring Calorie and Nutrition from Food Image. IEEE Transactions on Instrumentation & Measurement. 5. Meyers, A., Johnston, N., Rathod, V., Korattikara, A., Gorban, A., Silberman, N., et al. (2015). "Im2Calories: Towards an Automated Mobile Vision Food Diary." IEEE International Conference on Computer Vision (ICCV). 6. Chen, M., Yang, Y., Ho, C., Wang, S., Liu, S., Chang, E., & Yeh, C. (2012). "Automatic Chinese Food Identification and Quantity Estimation." SIGGRAPH Asia Technical Briefs. 7. Okamoto, K., & Yanai, K. (2016). "GrillCam: A Real-Time Eating Action Recognition System." International Conference on Multimedia Modelling (MMM). 8. Itakura, K., Narita, Y., Noaki, S., & Hosoi, F. (2021). "Automatic Pear and Apple Detection by Videos Using Deep Learning and a Kalman Filter." OSA Continuum. 9. Lawal, O. M. (2021). "YOLOMuskmelon: Quest for Fruit Detection Speed and Accuracy Using Deep Learning." IEEE Transactions on Agricultural Engineering. 10. Betzalel, N., Zehavi, A., & Ben-Zano, I. (2020). "Mobile-based Dietary Tracking Using Deep Learning and Image Processing." Journal of Nutrition Science & AI. 11. Zhang, S., Wang, Y., & Xu, J. (2021). "A Deep Learning Approach for Calorie Estimation Based on Food Images." Neural Computing and Applications. 12. Li, J., Dong, Q., & Tan, W. (2019). "Food Image Recognition Using CNNs for Automated Nutrient Estimation." Computer Vision and AI. 13. Wu, H., & Liu, X. (2020). "Advancements in Image-Based Dietary Monitoring Using AI." International Journal of Computer Vision. 14. Wang, X., & Song, Y. (2019). "Deep Learning in Food Classification: A Review." IEEE Transactions on AI. 15. Nguyen, T., Pham, K., & Tran, D. (2018). "A Cloud-Based System for Food Recognition and Calorie Calculation." Cloud Computing Applications. 16. Tada, T., & Hashimoto, Y. (2022). "Real-Time Dietary Analysis Using AI and Edge Computing." Journal of AI Applications. 17. Mohan, R., & Kumar, P. (2021). "Computer Vision Techniques for Nutrient Estimation." Machine Learning Journal. 18. Nakamura, M., & Yoshida, R. (2019). "Food Volume Estimation for Calorie Counting." International AI Symposium. 19. Singh, S., & Patel, V. (2020). "Enhancing Food Image Recognition with Transformer Models." Deep Learning Research. 20. Gupta, A., & Sharma, P. (2022). "AI-Driven Nutritional Analysis for Smart Health Applications." AI in Healthcare Journal.