Scribd
Upload
86 views12 pages

Calorie Detection and Alternate Food Recommendation System Using

1. The document proposes a calorie detection and food recommendation system using deep learning techniques to analyze images of food, predict calorie content, and suggest alternative foods if the calorie amount is too high or low. 2. It describes existing methods for food image classification, segmentation, volume and calorie estimation, and recommendation systems. The proposed system would use R-CNN for image classification, calculate calories from tables, and provide meal-specific recommendations based on calorie needs. 3. The system is intended to analyze images of mixed Indian foods using a dataset created for this purpose, in order to help users track calorie intake through automated image analysis and recommendations.

Uploaded by

swathi srini
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
86 views12 pages

Calorie Detection and Alternate Food Recommendation System Using

1. The document proposes a calorie detection and food recommendation system using deep learning techniques to analyze images of food, predict calorie content, and suggest alternative foods if the calorie amount is too high or low. 2. It describes existing methods for food image classification, segmentation, volume and calorie estimation, and recommendation systems. The proposed system would use R-CNN for image classification, calculate calories from tables, and provide meal-specific recommendations based on calorie needs. 3. The system is intended to analyze images of mixed Indian foods using a dataset created for this purpose, in order to help users track calorie intake through automated image analysis and recommendations.

Uploaded by

swathi srini
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 12

Calorie detection and alternate food

recommendation system using deep


learning techniques

S. Saran Swathi
Objective
• To detect the calorie intake of the food by capturing the image of the
food and also to recommend alternate food if the predicted calorie is
higher or lower than required.
Introduction
• In earlier models developed to predict the calorie intake of a person,
required manual input from the user. This method had a lot of issues.
• In order to simplify the method and to obtain the accurate intake
quantity, a new method came into existence in which people can
calculate the accurate calorie intake from the image of that particular
food.
• There are various algorithms available for object detection like R-CNN,
Fast R-CNN, Faster R-CNN and YOLO. Each method has its own pros
and cons.
• This paper analyses and compares all of these methods
Module Description:

1.Food images databases:


• High quality food pictures is essential to train a model for the
classification of food and standardize the performance of predicting.
The massive food image databases like Food-101, UEC Food-256 and
UEC Food-100 have been used in many papers. Some research papers
have developed their own database consisting of food items with
probe object. The probe object can be anything like coin or even a
thumb finger. The probe object is used to determine the quantity of
the food.
2.Food Classification:
• The Machine learning based approach for food image classification
consists of feature extraction and they are used to construct a
prediction model based on various algorithms like Support Vector
Machine (SVM), K Nearest Neighbours (KNN) and Bag of Features. The
emerging deep learning approach has a huge amount of connected
layers which learns various features and build a final layer which is
responsible for classification. Popular food classification approaches in
deep learning includes ANN, R-CNN, faster R-CNN. Deep learning based
approach is most popular since the accuracy obtained by this method is
very high when compared to the machine learning approach.
3. Food Segmentation:
• Image segmentation is the process of segmenting the image to
identify each object in the image. In order to identify each and every
food item present in the image, image segmentation is performed.
Image segmentation is very much important for mixed food items so
that every food item can be individually detected by the model.
 
4. Food volume estimation and calorie prediction:  
• After the identification of food images, the volume of the food in the
image has to be analysed in order to predict the amount of calorie in
the food. The volume estimation can be done by using the formula
and the calorie can be predicted by looking up the table.
5. Recommendation system
• Recommendation system is used to recommend something to the
user based on their previous likes, needs or requirement. Food
recommendation system is used to recommend food to the user if the
predicted calorie intake is higher or lower than the required amount.
K-means clustering is used for recommending the food.
Proposed System Overview:

• The proposed system consists of the following modules- i) Image


classification ii) Segmentation iii) Calorie prediction iv) Food
recommendation
• R-CNN is used for image classification and by looking up the table, the
calorie of the food is calculated.
• The food recommendation is based on the type of the meal. Since calorie
intake differs for morning, afternoon and night, the food that is
recommended will also very based on it. If the calorie predicted is same as
the required calorie intake, then there is no need of alternate food
recommendation.
• The dataset for proposed system consists of mixed Indian food items.
Overview of the proposed system
Dataset for proposed system
Conclusion

• Deep learning is playing a vital role in all domain particularly in health


care. With the help of image recognition the accuracy can be
enhanced and manual work is also minimized.
• In future, a model will be developed to predict the calorie and to
recommend an alternate food based on the predicted calorie.

You might also like