Image Recognition Native Mobile App

LocalIf ​​is an image recognition application specifically designed for IFES Campus Serra. With a comprehensive database, the program stores information about the various locations in the campus. Using two feature extractors, LocalIf ​​analyzes the photos sent by the user and determines the corresponding location. It also contains a pathfinder feature, which allows the user to find the shortest path between two locations.

Project Status

🟢 Functional: Access to functionalities through code or application.

Demo

Features and Application Demo

• The user can take pictures inside the campus using the Android compatible mobile app built with Kotlin in Android Studio, allowing convenient use at any time.
• Through image processing performed in Python, all photos in the database and photos taken by the user are converted to PGM format, standardizing them in grayscale.
• The algorithm is mainly implemented in C language, using two main structures for handling information.
• The first structure is responsible for creating the database, where the characteristics of each image are extracted and stored individually in text files (.txt).
• Descriptor extraction is performed using two different methods: histogram and statistical shape extractor. Extracted descriptors are saved in separate folders.
• The second framework is focused on image recognition. The input image descriptors are extracted and Euclidean distances are calculated between it and the database image descriptors.
• The results are averaged weighted on each base image, providing greater accuracy.
• The structure is sorted based on the results, from lowest to highest value, displaying the five most similar locations and their corresponding images. Note: The usage of pictures from locations other than IFES Campus Serra can lead to inaccurate results, as the database is specific to that place.

How to Use the C Algorithm using GCC

• The C algorithm is located in the image-recovery folder.
• The image-recovery folder contains the img folder, which contains the database images, and the test folder, which contains the test images.
• The image-recovery folder also contains the main.c file, which is the main file of the program.
• To run the program, you must have the gcc compiler installed on your computer.
• To compile the program, run the command gcc main.c -o query.
• To run the program, run the command ./query path/to/query/image.

Download the Android App

Download the app via Google Drive

Technologies Used

• C/C++
• Python
• Kotlin
• Java

Project Developers

• @Filipe Suhett Nogueira Silva
• @Giovanna Scalfoni Sales
• @Jhonata Polito Demuner
• @João Pedro Spinassé Viana
• @Marllon Christiani dos Santos Ribeiro