Automatic Video Processing for Body Worn Camera
15 Février 2023
Catégorie : Post-doctorant
It is a joint work between EURECOM (France) and UPNM (Malaysia).
Body worn cameras (BWCs) have been more popular over the last decade. They are becoming one of the essential tools for law enforcement officers to carry with them for surveillance purposes. The implementation of BWCs can be seen across Europe and the rest of the world. There are several types of BWCs utilized by law enforcement across the world. Some are attached to the officer's helmet, while others are linked to the officer's chest. Some record continuously, while others begin recording only when a certain trigger occurs, such as the opening of a car door or the activation of a siren. Generally, videos captured by BWCs are used a posteriori in case of major problems between police officers and citizens controlled (reactive surveillance). The main goal of this work is to develop video processing tools enablingproactive surveillance. Very limited studies exist in the domain. Because of the specificities of the BWCs, customized algorithms have to be designed and developed. Below the list of key elements and challenges to investigate:
- Cameras: They are different models with different capacities in terms of resolution, field of view, storage, wifi, infrared, etc. Existing technologies will be listed at the beginning of the project;
- Datasets: The number of existing datasets is very limited. If needed, a new dataset would be created;
- Quality evaluation of recoded videos in real time: it is important to be sure that recoding videos will be usable by humans and machines thank to an appropriate positioning (e.g. enabling face detection of person controlled) of the camera and appropriate set of acquisition parameters;
- Face recognition in adverse conditions: contrary to many other contexts (e.g. face recognition at borders), face images could be of low quality in terms of pose and illumination that could impose to revisit some existing algorithms in face recognition;
- Facial expression, micro expressions and emotion: extracting such information can provide cues to policemen concerning intention of the person controlled and his/her psychological status:
- Gait: It is also possible to extract information about the identity or behavior of a person by analyzing his/her gait. Existing studies have to be revisited to be adapted for BWCs.
Finally, it can be envisaged to identify the camera’s wearer from recorded videos.
EURECOM is a major Engineering School and a Research Center in digital sciences founded in 1991 as a consortium in the international technology park of Sophia Antipolis. The IMT is a founding member of the GIE. Teaching and research activities are organized around 3 promising fields: digital security, communication systems and Data Science.
EURECOM has a staff of 150 (researchers and support teams) and welcomes 400 international students on the Campus Sophia Tech, the largest information science and technology campus of the region. EURECOM enjoys a privileged geographical environment on the French Riviera (Côte d'Azur), between sea and mountains, at the heart of a dynamic and multidisciplinary ecosystem that promotes high-level scientific and technological innovation.
- Education Level / Degree : phd
- Field / specialty: Image processing / Computer Vision / Artificial Intelligence
The application must include:
- Detailed curriculum,
- Name and address of 2 references.
Applications should be submitted by e-mail to email@example.com the reference : DS/JLD/bodycam/022023