Visual Content Analysis
for Innovative Advertising and Selling

Stimulate purchases via any visual content
Make images and videos a showcase of goods and services.

Monetize visual content by stimulating purchases with moments of inspiration while watching videos.

Make advertising intelligent and effective through matching a visual content with relevant ads and offers.

Increase sales in a fashion e-commerce with search tools based on the appearance of goods.

Artificial Intelligence at the Core
Machine Learning, Computer Vision, Neural Networks
How it works:

Objects detection

Objects tracking in videos

Persons and face appearances detection

Gender recognition

Human pose estimation

Garments detection, segmentation and recognition

Similar offers search for any garment based on its neural network features

Video scene analysis

Video fragments classification and matching with categories of goods and services for smart advertising

Video sentiment recognition (in progress)

Ready for use digital AI-based platform for visual content processing

Shoppable Video Engine “Palantir”  
Make video an e-commerce platform
Accurate and fast (in real time) content recognition and fashion retrieval.

Ready-to-go pipeline to mark-up visual content.

Simple integration with a partner’s software using API.

The engine configurable for flexible and scalable using hardware, able working with GPUs of different types and distributed computing on servers and GPUs.

Contextual Advertising in Video
Place ads in videos intelligently and effectively
Match video content with relevant ads and offers.

An alternative to audience targeting. Target ads effectively without using cookies to protect user privacy.

Make advertising intelligent.

Remove the gap between the moment of inspiration and the actual purchase.

How it works:

  • Objects detection
  • Scene analysis
  • Video fragments classification
  • Video fragments matching with categories of goods and services to place relevant ads

The most common types of scenes in typical video content are recognized (for example, food, sports, pets, cafes).
Video stream processing in LIVE mode, with delay of maximum 1-2 seconds.
Video file offline processing mode for a time approximately equal to film duration.
Output data: start & end timestamps of a scene, type of content & recommended advertisement type.
Regulated frequency of recommended advertising points.
Human facial expressions and video sentiment recognition applied to determine content emotional appeal and its suitability and safety for brand advertising (in process).
Enhanced content monetization.
Cookie-less targeting.
Audience loyalty to contextual advertising: according to surveys, most consumers would be more likely to watch an ad if it matched content they are browsing.
Examples of successful contextual advertising campaigns on Russian TV channels:
• Sber Megamarket (advertising food delivery, pets keeping, and some household goods),
• Aquvue (advertising lenses when the viewer sees persons with glasses on the screen).

Fashion Visual Search Engine
Target via garments detected in images and videos
Fashion Visual Search Engine based on Computer Vision technologies automatically detects and recognizes garments shown in an image / a video scene and offers to buy similar products from the collections of e-commerce partners.

Fashion Engine allows impulsive purchases of goods seen in content.
How it works:
  • Object detection and tracking
  • Person appearance matching with one of video characters
  • Gender recognition
  • Pose estimation
  • Features extraction for garments
  • Similar offers search by look samples from video, based on features similarity of items

Image processing mode (suited for Instagram photos and the like), 1.2 sec. per image on average.
Video stream (from TV channels) online processing mode, with delay of maximum 2 seconds.
Video file offline processing mode, for a time approximately equal to film duration.
Able to work with millions of Stock Keeping Units in multiple shops, keeping the catalog up-to-date on the daily basis.
Enhanced content monetization.
Increased audience engagement.
Additional sales channel for fashion e-commerce partners.
“Palantir” version 1 was used in a third-party application to mark up clothing items for some regular shows on Russian TV Channels thus allowing TV viewers to access the merchandise in content, using the mobile app. and the catalog filled with goods from the Russian fashion market majors such as LaModa, Wildberries, and TSUM.

Solutions for Fashion e-Commerce
Increase sales through effective tools
based on product appearance analysis

Efficient alternative to manual tagging: number of tags per product is doubled with the similar accuracy.

Thousand tags supported: category and model of clothing, shades of colors, texture, trim, material properties, style, gender, special attributes such as sleeve shape, heel height, etc.

Providing meaningful titles in a readable form.


Search for similar offers by photo.

A product photo or any image uploaded by the user can be used as a search sample.

This efficient alternative to text search provides smart navigation in online fashion shops.

All detected garments in the uploaded photo are samples for retrieval enabling the user to reproduce the whole seen look.


Pick up variants of complementary clothing products to combine them with the sample item or with its analogues.

Scalable solution: no manual marking required.

Cross-selling are stimulated.


Get the image roughly predicting how the user will look in the selected dress.

Two images required: photo of the desired dress, demonstrated by the model in online shop, and user’s photo.

Give customers the opportunity to see their image in the selected garments and to make informed decision about purchase.

Remove barriers to online shopping
and reduce returns.

The approach is based on the Generative Adversarial Network (GAN).


Boost sales by showing products that match consumer preferences
in style, shape, color, pattern, ets.

Prevent useless showing products that the consumer will never buy.

Audience loyalty and engagement growth.
Conversion Rate uplift.
Increasing the average check thanks to cross-selling.
Identification of popular fashion trends based on image recognition.
Trusted by Russian market majors such as KupiVIP, OZON, Joom.


CV expert in visual content understanding specializing in fashion recognition,

powering intelligent advertising and innovative sales practices

Our team has become the top winner of the clothing recognition competition:
iMaterialist Challenge at FGVC 2017

High recognition accuracy and real-time performance characterizes DressCoder products and solutions.

Get in Touch

Please contact us at

This site was made on Tilda — a website builder that helps to create a website without any code
Create a website