Ball Detection and Tracking Model using TensorFlow TFLite

Project Description:

I am looking for a skilled developer to implement a ball detection and tracking model using TensorFlow TFLite on the Android platform. The model should be based on Recurrent Neural Networks (RNN) and trained using image data.

Ideal Skills and Experience:

Task Title: Ball Detection and Tracking Model using TensorFlow Lite

Task Description:

We need a TensorFlow Lite (TFLite) model with the ability to detect and track balls. If you already have a TFLite model file that meets the requirements, you can directly replace the existing assets/models/[login to view URL] file to observe its performance. The attached APK provides a preview of the expected result. The model should solely focus on detecting balls and draw bounding boxes around them, without detecting any other objects.

Task Requirements:

Train a TensorFlow model to accurately detect and track balls.

Export the trained model in TFLite format for efficient execution on mobile devices.

If you already have a pre-trained TFLite model that meets the requirements, replace the specified file path (assets/models/[login to view URL]) and ensure that it fulfills the task requirements.

Provide a demonstration APK containing the ball detection and tracking model you trained or integrated. The APK preview should demonstrate accurate ball detection and bounding box visualization, without misclassifying other objects as balls.

Additional Information:

To ensure efficient execution on various mobile devices, make sure the model has low computational and memory overhead. We expect the model to run smoothly in real-time scenarios.

Please provide documentation with the submission, including the steps for model training, dataset description (if applicable), model performance evaluation, and any specific considerations required.

We are looking forward to collaborating with experienced candidates who have a deep understanding of image recognition and TensorFlow. Together, let's accomplish the task of creating a reliable ball detection and tracking model. If you have any questions regarding the task, feel free to reach out to us.

Attached Image: (sample images demonstrating ball detection for reference.)

- Proficiency in TensorFlow and TensorFlow Lite

- Experience with developing applications on the Android platform

- Strong understanding of Recurrent Neural Networks (RNN)

- Knowledge of computer vision and object detection techniques

- Familiarity with image data processing and training models using image data

Python Tensorflow Flutter

ID projekta: #36934897

Več o projektu