Skip to content


Our AI solutions related to Neural search has been developed in-house for the specific objective of applying deep learning technology to the fashion industry. We have conducted an extensive search on million product images, a large scale search on a single CPU processor with an acceptable latency. We build a bot that automatically extracts product from a thousand of fashion websites in Vietnam, a magnitude order of terabyte data.

The concept of semantic text search is to represent both products and queries as a semantic vectors in the multi-dimensional semantic vector space. Products and queries have to be mapped to vectors in such a way that similar products and queries close by meaning would be clustered together. This is achieved by training a deep learning model based on combining pattern, texture, time, space to find the best possible vector representation.

Visual search is literally finding similar images by similar vectors, this search engine converts an image into an embedding vector, and it traverses through the index to find other embedding vectors that are the most similar.

Dazlinn’s AI models perform all tasks required for efficient visual search, from object detection to vector representation and vector search of the target image.

Data extracting automatically

To simplify the working process, reduce the workload and time for clients, Dazlinn has built the crawling system that automatically collects all the needed data from client’s website.