.png){kind=small}
Create Your First Project
Start adding your projects to your portfolio. Click on "Manage Projects" to get started
Automatic Consumer Goods Classification Engine
Project type
Classification
Tools
Python, OpenCV, TensorFlow/Keras, Scikit-learn, NLTK/Spacy, Matplotlib, Seaborn, Pandas, NumPy, SIFT, ORB, SURF, Pre-trained CNN Models (e.g., VGG16, ResNet), Bag-of-Words, Tf-idf, Word2Vec, GloVe, FastText, BERT, Universal Sentence Encoder (USE)
Skills
Data preprocessing, Feature extraction, Dimensionality reduction, Machine learning, Model evaluation, Data visualization, Communication skills, Cloud services knowledge, Version control systems knowledge
Conducted a feasibility study for developing an automatic classification engine that categorizes consumer goods using both textual descriptions and image data.
Performed comprehensive data preprocessing, including cleaning and preparing textual descriptions, and processing images for feature extraction. Extracted meaningful features using various methods: implemented CNN transfer learning for image features and utilized multiple text processing techniques like Bag-of-Words, Word/Sentence embeddings (Word2Vec, BERT, and Universal Sentence Encoder).
Applied dimensionality reduction to project these features into a 2D space for visualization and analysis, assessing whether visual and textual data could effectively distinguish product categories. Validated clustering results by comparing them with actual categories to verify feasibility.
