What Is Visual Search?
Visual Search, sometimes known as Reverse Image Search, is the ability to search by image in a number of different ways. This can either be enhanced with text, or as a standalone image match or photo search function. In the world of retail this is particularly useful as it enables the customer to find exactly what they are looking for in a quick and convenient way.
Using apparel as an example, trying to find a particular ‘look’ may be difficult to describe, but being able to search by image allows the customer to find something relevant - either an exact match or something similar. This enables retailers to provide their own product inventory when a customer does a reverse image search. Typically a customer may use a photo taken on a mobile device or an image found somewhere on the internet / social media. Our technology allows users to gain inspiration from any number of image sources.
Why We Are World Leading
Developed in Imperial College London, Cortexica’s unique technology is based on human visual processes. It mimics how the brain’s visual cortex perceives images, rather than purely machine-based learning.
Here are a few reasons why Cortexica is a global leader at delivering image recognition and other solutions to search by image:
Cortexica delivers accuracy of 95%, the same as a human being.
Our image results for clients consistently deliver response times of less than one second – around 10 times faster than other visual search technologies such as Mechanical Turk.
Our flexible, cloud-based infrastructure allows us to instantly scale new hardware to deliver when clients hit peak hours. We use the best technology available, utilising fast GPU servers to meet the needs of client demands.
Alongside cutting edge-technology Cortexica has an experienced team of retail experts to develop effective photo search and image match retail solutions for clients.
Our team also includes a number of PhD research scientists and engineers who research and tailor solutions to meet the needs of our clients, combining quality and speed. We have a comprehensive toolbox that can be utilised across any platform.
About Our Technology
In 2005, Cortexica’s patented technology was conceived in the labs of Imperial College London by world renowned founder Dr Anil Anthony Bharath and Jeffrey Ng. Over a decade has passed and the technology has matured into the best reverse image search solution for retail in the world.
Our 30 strong R&D scientists, which includes 9 PhDs, have developed and enhanced our solutions utilising our very own “secret sauce”. Computer Vision and Deep Learning (supervised and unsupervised) techniques form the basis of our technology and provide the ability to search by image with speed and accuracy.
Using our own code base means we can combine and adjust configurations to meet the demands of individual clients. This customisable approach is delivered using fully automated RESTful APIs that our clients can plug into their software, apps and websites.
Our core technology:
Fundamentally, our technology provides the capability to search by image to find similar or matching images and product items. We work with our clients to understand each use case and combine our core technologies to provide the highest accuracy and fastest response times – under 1 second from a customer uploading a new image to presenting your inventory results.
Automated - Our technologies can automatically segment and detect your type of image, apply the right configurations and automatically process a response in less than one second.
Computer Vision Methods - Our patented methods process several visual channels in parallel, such as scale, texture, colour and shape, looking at 1500 points of interest in each image. Depending on our client’s industry we can utilise other techniques. For example, within fashion we can use face/body detection to deliver higher accuracy.
Deep Learning (supervised and unsupervised) - Depending on requirements we utilise different models (RNN & CNN) to learn, detect and categorise. We can even base our models on our clients' own unique images and metadata, as we firmly believe that one size does not fit all. We then automatically identify and categorise any of our client’s image assets and auto tag their products.
Applying a mix of our core technologies, we deliver numerous solutions, such as Product Discovery – for Exact Match and findSimilar, Duplicate Detection, Auto Categorisation, Auto Tagging, Brand Monitoring (Video) and Product Recommendation.
Our technology is delivered over an integrated, secure and scalable cloud service and is built to easily scale to our clients’ needs.
Our service is running over the latest available high performance NVIDIA GPUs to provide the very best results in accuracy and speed.
We can scale our service to billions of images whilst maintaining the highest speed and accuracy.
Your business isn’t in one location; we can provide a distributed solution which means your customer, wherever they are in the world, will get the very best response times.
Secure and Reliable:
In designing our service, we made sure there is complete redundancy and we continuously secure the system from threats.
Our findSimilar core technology brings together years of research and experience in the retail industry. Using Computer Vision and Deep Learning techniques we provide image search solutions that deliver measurable results to our clients.
At the simplest level, our solutions allow customers to search by image for similar items in a designated product inventory. Used primarily for e-commerce, this can significantly increase conversion rates and basket size by providing relevant product results.
Reverse Image Search
Utilising our Exact Match core technology we also provide solutions that can be used to reverse image search to find matching products and logos. Analysing attributes such as pattern and shape, our technology can scan images and videos to find image match items, even if they are rotated or skewed.
In addition we provide a number of solutions to recognise duplicates and even product recognition within video content – with each solution customised to the client’s specific requirements in every case.