Usama Zafar

Academic

Career

Verbatim text

FruitVegFreshness 
FruitVegFreshness is an android application that allows the users to take images of fruits and 
vegetables either by capturing the photo directly by using the mobile camera or by loading it from 
the mobile gallery. These images are then forwarded to the remote server on which the object 
detection model is already deployed, where the model analyzes, locates the objects in the images 
by drawing bounding boxes around them and predicts the freshness level (fresh, medium or rotten) 
along with guessing the fruit/vegetable type like apple, banana etc. Finally, after the prediction has 
been made, the server returns the image with predictions that contains bounding boxes around 
objects along with labels, to the client user. This resultant image with predictions is finally 
presented on the screen of user’s device.  
Features include: 
 
Interactive UI for taking picture using Camera and loading of image from Gallery. 
 
Accurate and precise YOLOv5 Object Detection model deployed on server. 
 
Localization of fruits and vegetables in the image by drawing bounding boxes around them. 
 
Classification of freshness and type of fruits and vegetables present in the image like fresh 
apple, rotten apple etc. 
Zooming in and out from the resultant images with predicted bounding boxes and labels. 
 
 
 
 
 
 
 
 
 
Technology Used: 
Python, Java, Android Studio,  
Google Colab & PyTorch 
Supervisor Name: 
Dr. Labiba Fahad 
Group Members:   
Usama Zafar (i17 - 0012) 
Usama Rasheed (i17 - 0212) 
Hafsa Saqib (i17 - 0321)

AI enrichment