Welcome TO IDASO RESEARCH & DEVELOPMENT
Idaso is an innovative pioneer in the area of data analysis, and computer vision research and development. We help organisations, from government authorities down to young start-ups, understand and improve their businesses by leveraging the inherent value of their data. We work with our clients to encourage adaption of Video Intelligence analysis to gather insightful facts, figures, movements and patterns relative to their customer’s needs. The results from the analysis gives our clients the benefit of putting logical implementation plans and actionable infrastructure strategies in place to improve plans, processes, operational inefficiencies or future services for continued growth. If your company or organisation have a problem and are interested in collaborating to find answers, let us put our R & D expertise into action and refine a solution that meets your needs. Lets put our technology at work for you!
ABOUT idaso r&d
The Core Competency of our Software Development team is the training of customised Neural Networks to extract "Objects of Interest" from images be they still frames or from CCTV. This enables the conversion of CCTV images into actionable intelligence. Our solutions are industry agnostic, we simply customise our systems to suit the needs of our clients and partners. Our specialities are driven in the use of Artificial Intelligence powered image processing to extract data from video files. Computer Vision is the science and technology of making machines that see. It is concerned with the theory, design and implementation of algorithms that can automatically process visual data to recognise objects, track and recover their shape and spatial layout. Idaso has developed their own bespoke unique algorithmic approach to tracking and identifying objects of interest and moving objects from CCTV. Coupled with a robust self-diagnosing accuracy system this approach has enabled us to automate much of the manual processes involved in collating qualitative data from CCTV. The use cases of our bespoke technology has been very effective in Vehicle Movement Data Analytics/Collection support services, particularly for the traffic and transportation industries to date.
5 REASONS TO CHOOSE IDASO
Reason #1: We focus on building partnerships!
Partnerships are everything to Idaso, and it shows in everything we do. We’re sympathetic to your challenges and excited by your achievements and opportunities. We think you’ll find us professional, friendly and supportive.
Reason #2: We can take care of all your data analysis requirements!
We have a talented and experienced team with exceptional skill levels, from development to implementation. Specialising in machine vision learning algorithms, we are dedicated to matchmaking your business with the right technology for future efficiency.
Reason #3: We provide ongoing support!
As an integral part of our dedicated service, we provide ongoing support and maintenance to all our clients. Rest assured whatever your particular query or issue may be, we are here to support you and to render a solution with the minimum of fuss.
Reason #4: We have a proven reputation!
We have delivered successful projects to a broad range of clients, from start-ups to small and medium sized businesses, publicly listed companies and governmental sector. We have an extensive portfolio of past projects, and successful service solutions expanding every day.
Reason #5: We love what we do!
Sharing our wisdom and creating new opportunities for businesses is what challenges and excites us. We realize that every client is unique and we work hard to materialize their vision and goals. The difference is in the detail.
innovation & CONTINUOUS Development PROJECTS
Idaso’s particular strength’s apart from standard computer vision and vision based artificial intelligence solutions (Convolutional Neural Networks, CNNs) continue to evolve every day. Below are some examples of our in-house innovative research and development team’s recent projects that have delivered customer value and/or significant efficiency gains in development activities:
- CNN based object of interest detection's, presentation of detection results on a heat-map. The heat-map sets the spacial frequency whereabouts of the objects of interest, e.g. locations where cars are usually parking, people moving, etc...
- Object tracking on live scenes or captured video, using various object tracking algorithms optimised for speed and/or accuracy.
- Automatic clustering of tracks of objects to facilitate the identification of irregular tracks in the spacial and/or temporal domain. This technique is used for error detection or security purposes.
- Extraction of synthetic training data from widely used computer games. By successful resolution of this technically very challenging task we drastically reduced our CNN training data generation costs.
- Enrichment of existing training data by synthetic noise (e.g. motion blur) and light/whether conditions (e.g. dawn/dust/rain/snowfall). This technique helps us to reduce costs and speed up implementations: we simulate visual conditions where object detections must perform yet gathering real training data is expensive/slow.
- Idaso have implemented a CNN based object detection and object tracking methodology in Azure cloud environment. Thus we are able to scale up data processing to practically any demand.
- Idaso have also developed a user interface application to interactively analyse tracks of objects on captured videos. The API speeds up the analysis of long duration videos by two orders of magnitude. This API can also be used for security applications as is has various search filters for tracks.
Our trained neural networks can summarise long video files to the key moments that you, as an observer are interested in, thereby saving thousands of man hours. This is an industry challenging problem because finding important or informative parts of the original video requires to understand its content. Furthermore the content of Internet videos is very diverse, which makes video summarization much more tough as prior knowledge is almost not available. To tackle this problem, using deep video features that can encode various levels of content semantics, including objects, actions, and scenes, improving the efficiency of standard video summarization techniques. Using design a deep neural network or CNN’S to generate a video summary, we extract the deep features from each segment of the original video and apply a clustering-based summarization technique to them. The results demonstrate the advantages of incorporating our deep semantic features in a video summarization technique for future client services and needs.