Fruit Identification and Nutritional Analysis of Indigenous Indian Fruits Using MobileNetV2 emphasizing Vitamin C and Potassium

Image identification has been revolutionized by progressions in computer vision and deep learning. Nutritional analysis, accurate identification, examination of twelve distinctive indigenous fruits by an architecture known as MobileNetV2: fig, guava, pineapple, cherry, mango, tamarillo, apple, banana, mulberry, carambula, are the objectives of this research. MobileNetV2 is particularly relevant for deployment on resource-limited devices, which is applicable in real-time applications in the food and agriculture sectors, due to its computational efficiency. Our model uses transfer learning from the ImageNet database in order to accurately classify such fruits highlighting extraction of nutritional information mainly vitamin C & potassium. The model accuracy had a big improvement during training with training loss decreasing from 1.2984 to 0.0482 across 20 epochs while accuracy went up to 98.44%. For consumers who want healthy eating habits, this tool provides an easy way access information on nutrition labels found on various fruits which goes a long way in helping them choose what they want at the same time in the world of foods things have got easier because fruit recognition has been automated thus maintaining uniformity during labeling as well as accuracy regarding nutrition facts.