Applied AI Lab: Deep Learning for Computer Vision
Issued by
WorldQuant University
Earners of this badge completed six end-to-end PyTorch computer vision projects. From data preparation, cleaning and transformation pipelines to mastering deep learning models like MLPs, CNNs, and transformers, they tackled tasks like image classification, object detection, and generative AI. They applied transfer learning and self-regulated learning, explored software libraries, and debugged code while considering core values and ethical and environmental concerns faced by AI scientists.
- Type Learning
- Level Advanced
- Time Weeks
- Cost Free
Skills
- Artificial Intelligence (AI)
- Data Science
- Deep Learning
- Facenet
- Facial Recognition
- Flask
- Generative Adversarial Networks (GANs)
- HuggingFace diffusers
- HuggingFace Hub
- HuggingFace transformers
- Image Classification
- Image Generation
- Machine Learning
- Matplotlib
- MediGan
- Neural Networks
- NumPy
- Object Detection
- OpenCV
- Pandas
- Python
- PyTorch
- Scikit-learn
- Stable Diffusion
- Supervised Learning
- Ultralytics YOLOv8
Earning Criteria
-
Learners completed six end-to-end projects. Each project consists of four self-paced lessons, followed by an assignment that is programmatically graded.
-
1. WILDLIFE CONSERVATION IN CÔTE D'IVOIRE: In this project, learners look at a data science competition helping scientists track animals in a wildlife preserve. The goal is to take images from camera traps and classify which animal, if any, is present. To complete the competition, learners expand their machine learning skills by creating more powerful neural network models that can take images as inputs and classify them into one of multiple categories.
-
2. CROP DISEASE IN UGANDA: In this project learners work with a dataset of crop disease images from Uganda. They build and train a convolutional neural network to classify images into five categories. In the project they learn how to improve the performance of a computer vision model by using pre-trained models and by optimizing training with techniques like Callbacks.
-
3. TRAFFIC MONITORING IN BANGLADESH: In this project, learners work with traffic video feed data from Dhaka, Bangladesh. The goal is to take each frame of video and detect and label objects such as cars and people in real-time. They use both a pre-trained model but extend an existing model to detect custom objects for the task of analyzing traffic feed data.
-
4. CELEBRITY SIGHTINGS IN INDIA: Learners perform face detection and recognition using a video of boxer Mary Kom. They use pre-trained MTCNN and Inception-ResNet models to create face embeddings for Mary Kom and her interviewer, enabling face detection in new images. The project concludes with wrapping their code into a Flask app, allowing users to upload an image and perform face recognition.
-
5. MEDICAL DATA GENERATION IN SPAIN: In this project, learners use neural networks to generate new images. They do this using a Generative Adversarial Network (GAN) system, both by building one and using a pre-trained GAN. The images they generate will be a variety of medical images, such as X-rays and MRIs. They also create a web app using Streamlit to allow users to interact with the GAN. Additionally, they use Git and GitHub to track the app's code.
-
6. SOCIAL MEDIA MARKETING IN THE UNITED STATES: Learners a stable diffusion model to create and deploy a meme generator app on Streamlit for social media marketing in the United States.