Visual Content Analysis
for E-commerce and Advertisement Markets

Artificial Intelligence tools for sales growth

A high-performance platform based on Neural Networks
is commercially used in several massive scenarios:

1
Video Content Targeting Engine “Palantir”

makes video a showcase of goods and services

Content-based advertising as a way of additional video monetization in the era of extended privacy protection standards
This video demonstrates how technologies help to use the video content as a context for displaying ads:

  • Features
    Accurate and fast (in real time) visual content understanding, fashion recognition and fashion retrieval.

    Configurable and flexible framework with support of hardware-based scalability.

    Able working with GPUs of different types.

    Distributed computing on servers and GPUs.

    Ready-to-go pipeline.

    API for simple integration.

  • Benefits
    Provides a new sales channel through the ability to handle impulsive purchases
    Greatly increases the size of advertising properties
    Improves engagement and retention by making ads less intrusive (comparing to mid-rolls and banners with behavior-based targeting)
    Creates targeting possibilities in non-personalized channels (i.e. linear TV) via high-performance, real- time, automated, cost-effective labelling
    Restores ads efficiency in the world of new privacy standards in the Internet
    Enables a whole new array of services based on intelligent convergence between e-comm and media
  • Trusted

    The engine was used to search for analogues of clothes from TV shows (in third-party app) and for contextual ad on HbbTV.  


1a
Contextual Advertising in Video
Match video content with relevant ads and offers

How it works:

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

  • Special Features

    The most common types of scenes in typical video content are recognized and associated with ad categories, for example food delivery, sporting goods, pets keeping, and café services.

    High-accuracy scene type recognition.


1b
Fashion Recommendations from Video
Search for products similar to clothes from the video
How it works:
  • Object detection and tracking
  • Person appearance matching with one of video characters
  • Gender recognition
  • Pose estimation
  • Clothes recognition
  • Features extraction for garments
  • Similar clothing items search by video, based on features similarity

  • Special Features

    Variety of supported object types: clothes, shoes, hats, bags, and accessories, such as ties and glasses.

    Able to work with millions of Stock Keeping Units in multiple shops, keeping the catalog up-to date on a daily basis.

    High-accuracy recommendations of similar items.


2
Solutions for Fashion e-Commerce
Increase sales through efficient tools
based on product appearance analysis
AUTOMATIC TAGGING AND TITLES GENERATION

AI tagging is an efficient alternative to manual tagging: number of tags per product is doubled with 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.


COLOR SHADES RECOGNITION

140 color shades are recognized and can be used for filters and textual search.




SEARCH FOR SIMILAR CLOTHES

Similar offers search by image.

A product photo or any image or video 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 can be used as inputs for retrieval enabling the user to recreate the whole outfit.



OUTFITTING RECOMMENDATIONS /
SEARCH FOR COMPLEMENTARY ITEMS

Complementary items search by image of reference clothes.

Generated combinations in clothes are perceived as advices from a human stylist.

Scalable solution: no manual labeling required.

Cross-selling as a benefit.



VIRTUAL TRY-ON
(in progress)

Generate an image roughly predicting how the user would look in the selected dress.

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

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

Remove barriers to online shopping
and reduce returns.



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

  • Features

    Variety of supported object types: clothes, shoes, hats, bags, and accessories, such as ties and glasses.

    Variety of image types: product photos, images from the Internet, UGC, video.

    High-accuracy recommendations of similar and complementary items.

    Real time processing.

    Configurable and flexible framework with support of hardware-based scalability.

    Able working with GPUs of different types.

    Distributed computing on servers and GPUs.

    Ready-to-go pipeline.

    API for simple integration.

  • Benefits
    Online shopping audience loyalty and engagement growth.
    Cross-selling.
    Sales and retention growth.
    Identification of popular fashion trends based on image recognition.
  • Trusted
    Trusted by independent companies including two large European marketplaces.

Technologies
Machine Learning, Computer Vision,
Neural Networks, Deep Learning
How it works:

Object detection

Object tracking in videos

Detection of a person’s appearance

Face detection

Gender recognition

Human pose estimation

Clothes detection, segmentation and recognition

Visual search for clothes (image search / video search): similar clothing retrieval for any garment by its image based on neural network features

Video scene analysis and recognition

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

Video sentiment recognition (in progress)

Ready for use digital platform for image/video processing


DressCoder

CV expert in visual content recognition and Content-based Image Retrieval

specializing in fashion recognition,

with a focus on powering intelligent advertising and innovative sales systems

Award-winning team:

- First place in the clothing recognition competition: iMaterialist Challenge at FGVC 2017

- Two patents in Computer Vision

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


Get in Touch

Please contact us at dresscoder.team@gmail.com

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