Computer Vision News - October 2016

Our CEO Ron Soferman has launched a series of lectures to provide a robust yet simple overview of how to ensure that computer vision projects respect goals, budget and deadlines. This month, we will learn about the Full kit of Requirements . One of the principles in project and software development is the need for a full kit of requirements before starting the actual development. It may happen that the image processing team is given many tasks with tight deadlines, in order to support the roadmap of product development. This often leads the process into a bottleneck. If the project is not launched in a systematic way, it might take several development cycles until the desired product is complete. In many cases, the reason is that the full specifications of the product are not well defined since the beginning: the development team is eager to start its work, managers are very determined to keep up with schedule and everyone wants to start as soon as possible. But without a full kit including all the detailed requirements which are needed for the development, deadlines can’t be met. Full kit for research: Sometimes, project managers sense that full requirements cannot be defined upfront because of technical and algorithmic unknowns. General product requirements are probably known, but they may be fluctuating until the software development step. Nevertheless, I recommend that the full kit in the research stage include all that is required for research: complete and realistic data in sufficient quantity, without which the developed algorithm might behave differently when confronted with real data. The research dataset should include all scenarios, especially in the medical field where an algorithm tested on data belonging to healthy people might be useless when used on patients with pathologies. Collecting relevant data and analysing its possible appearances is a difficult but necessary task. The reason for the need of a full kit in research is that algorithm must take all cases into account. With only limited examples, it will be a naïf algorithm that will not hold when used on the full variety. “ Without a full kit including all the detailed requirements, it might take several development cycles until the product is complete ” 34 Computer Vision News Project Management Tip Full-kit of Requirements Management “ Include complete and realistic data in sufficient quantity ”