Limits...
Tool for Semiautomatic Labeling of Moving Objects in Video Sequences: TSLAB.

Cuevas C, Yáñez EM, García N - Sensors (Basel) (2015)

Bottom Line: An advanced and user-friendly tool for fast labeling of moving objects captured with surveillance sensors is proposed, which is available to the public.The labeling can be performed easily and quickly thanks to a very friendly graphical user interface that allows one to automatize many common operations.This interface also includes some semiautomatic advanced tools that simplify the labeling tasks and drastically reduce the time required to obtain high-quality results.

View Article: PubMed Central - PubMed

Affiliation: Grupo de Tratamiento de Imágenes, Universidad Politécnica de Madrid (UPM), E-28040 Madrid, Spain. ccr@gti.ssr.upm.es.

ABSTRACT
An advanced and user-friendly tool for fast labeling of moving objects captured with surveillance sensors is proposed, which is available to the public. This tool allows the creation of three kinds of labels: moving objects, shadows and occlusions. These labels are created at both the pixel level and object level, which makes them suitable to assess the quality of both moving object detection strategies and tracking algorithms. The labeling can be performed easily and quickly thanks to a very friendly graphical user interface that allows one to automatize many common operations. This interface also includes some semiautomatic advanced tools that simplify the labeling tasks and drastically reduce the time required to obtain high-quality results.

No MeSH data available.


Related in: MedlinePlus

Mean time values (seconds) expended on the performed experiments.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4541825&req=5

f13-sensors-15-15159: Mean time values (seconds) expended on the performed experiments.

Mentions: Figure 13 illustrates the mean time values spent on each test and for each image. It can be observed that all of the advanced tools (Tests 2–5) give rise to a reduction of the time required to label (compared to the time spent on Test 1, in which only manual tools are used). In Test 2, where the labeling starts from the contours obtained for previous images, a significant reduction has been obtained for both images. This reduction is higher for Image B, since in this image, the object barely moves. The lowest computational reduction has been obtained in Test 3. This is due to using a motion detection mask, so that it is very difficult to obtain an initial contour with enough accuracy, since the edges of the masks are frequently inaccurate and many reflects cast by the moving objects are usually included as part of the detection. Consequently, users must apply many corrections to obtain the final labeling. However, as is seen in Figure 13, in Test 4, where the motion detection tool is combined with the use of the intelligent pencil tool, an additional reduction of the time required for the labeling has been achieved. This is because the intelligent pencil allows one to easily select the valid areas from the motion detection mask and, simultaneously, to draw accurate contours in those areas where the motion detection fails. Finally, it can also be observed that the active contours tool (Test 5) provides the best results (the lowest computational costs) in both images, since it applies a very fast algorithm that is able to provide a very good approximation to the labeling targets. Therefore, very few changes must be done by the users after using this tool.


Tool for Semiautomatic Labeling of Moving Objects in Video Sequences: TSLAB.

Cuevas C, Yáñez EM, García N - Sensors (Basel) (2015)

Mean time values (seconds) expended on the performed experiments.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4541825&req=5

f13-sensors-15-15159: Mean time values (seconds) expended on the performed experiments.
Mentions: Figure 13 illustrates the mean time values spent on each test and for each image. It can be observed that all of the advanced tools (Tests 2–5) give rise to a reduction of the time required to label (compared to the time spent on Test 1, in which only manual tools are used). In Test 2, where the labeling starts from the contours obtained for previous images, a significant reduction has been obtained for both images. This reduction is higher for Image B, since in this image, the object barely moves. The lowest computational reduction has been obtained in Test 3. This is due to using a motion detection mask, so that it is very difficult to obtain an initial contour with enough accuracy, since the edges of the masks are frequently inaccurate and many reflects cast by the moving objects are usually included as part of the detection. Consequently, users must apply many corrections to obtain the final labeling. However, as is seen in Figure 13, in Test 4, where the motion detection tool is combined with the use of the intelligent pencil tool, an additional reduction of the time required for the labeling has been achieved. This is because the intelligent pencil allows one to easily select the valid areas from the motion detection mask and, simultaneously, to draw accurate contours in those areas where the motion detection fails. Finally, it can also be observed that the active contours tool (Test 5) provides the best results (the lowest computational costs) in both images, since it applies a very fast algorithm that is able to provide a very good approximation to the labeling targets. Therefore, very few changes must be done by the users after using this tool.

Bottom Line: An advanced and user-friendly tool for fast labeling of moving objects captured with surveillance sensors is proposed, which is available to the public.The labeling can be performed easily and quickly thanks to a very friendly graphical user interface that allows one to automatize many common operations.This interface also includes some semiautomatic advanced tools that simplify the labeling tasks and drastically reduce the time required to obtain high-quality results.

View Article: PubMed Central - PubMed

Affiliation: Grupo de Tratamiento de Imágenes, Universidad Politécnica de Madrid (UPM), E-28040 Madrid, Spain. ccr@gti.ssr.upm.es.

ABSTRACT
An advanced and user-friendly tool for fast labeling of moving objects captured with surveillance sensors is proposed, which is available to the public. This tool allows the creation of three kinds of labels: moving objects, shadows and occlusions. These labels are created at both the pixel level and object level, which makes them suitable to assess the quality of both moving object detection strategies and tracking algorithms. The labeling can be performed easily and quickly thanks to a very friendly graphical user interface that allows one to automatize many common operations. This interface also includes some semiautomatic advanced tools that simplify the labeling tasks and drastically reduce the time required to obtain high-quality results.

No MeSH data available.


Related in: MedlinePlus