Final Increment - acvc279/Python-Project Wiki

OBJECT DETECTION

TEAM MEMBERS:

  • China Venkat Chowdary Arikatla
  • Venkata Rahul Torlikonda

The story of the project

Now a days, there are a number of disabled people, specifically, partial vision around the world needs help through using available technology. Partial vision occurs when a human unable to see normally or partial loss of vision. In stated that, most of partial vision people have come across embarrassing moments in between societies due to their lack of vision. And this loss is making their daily life harder than normal human beings. In this project, the aim is to use available technologies to create and develop a way to help them with detecting some objects throughout their daily activities. Machine Learning is the approach of this project.

The data that helps to develop the project

The dataset is about object detection. Fruits are in the dataset which we considered as an objects. That represents four kinds of labels to detect. In these project. We use a Fruits dataset which have around 460 train files and about 120 test files along with .csv files in it. .Csv file has following labels: apple, banana, mixed, orange. We extracted these dataset from kaggle.

Improment from previous increment

In previous increment we use pretrained coco dataset and also we detect object by using opencv. But in these increment we extracted a fruits dataset from kaggle and we implement without using opencv. Instead of opencv we used pygame to take an image from camera to detect the object.

Important code snippets from the project

  • Imported all the required libraries
  • Processing the training data
  • Processing the testing data
  • Visualizing the training data
  • Creating a Sequential model
  • Train the model using fit function
  • Plot of accuracy and loss
  • Testing prediction
  • Loading the image from camera
  • Input:
  • Output:

Work sharing/Module sharing between teammates

China Venkat Chowdary Arikatla: Processing the train and test detase. Implement pygame for prediction Venkata Rahul Tolikonda: Extracted the dataset from kaggle. Implemented the convelution neural network model on it.

Challenges Faced

Reshaping the inappropriate dimensions of an images in dataset Replacing the opencv with pygame

References