Open Access

Backpack detection model using multi-scale superpixel and body-part segmentation

Rahmad Hidayat
Rahmad Hidayat
, 
Agus Harjoko
Agus Harjoko
 and 
Aina Musdholifah
Aina Musdholifah
   | Aug 19, 2023
International Journal on Smart Sensing and Intelligent Systems's Cover Image
About this article
Previous Article
Next Article
Cite
Article Category: Article
Published Online: Aug 19, 2023
Page range: -
Received: Oct 09, 2021
DOI: https://doi.org/10.2478/ijssis-2023-0008
Keywords
© 2023 Rahmad Hidayat et al., published by Sciendo
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Figure 1:

Backpack detection using multi-scale superpixel segmentation and body-part method.
Backpack detection using multi-scale superpixel segmentation and body-part method.

Figure 2:

The resulting image of the foreground detection process on each dataset.
The resulting image of the foreground detection process on each dataset.

Figure 3:

Cell and block in 64 × 128 image.
Cell and block in 64 × 128 image.

Figure 4:

Human Body Proportion Model [29].
Human Body Proportion Model [29].

Figure 5:

Heads segment sample.
Heads segment sample.

Figure 6:

Bend-line identification and superpixel selection process.
Bend-line identification and superpixel selection process.

Figure 7:

Camera configuration in the acquisition room.
Camera configuration in the acquisition room.

Figure 8:

The segmentation results on the l1, l2, and l3 scales.
The segmentation results on the l1, l2, and l3 scales.

Figure 9:

The result of bend-line determining process on image with scale l2.
The result of bend-line determining process on image with scale l2.

Figure 10:

Example of Features Extraction on Selected Superpixels (B+h).
Example of Features Extraction on Selected Superpixels (B+h).

Figure 11:

The ROC curve for each scenario on the DIKE20 dataset.
The ROC curve for each scenario on the DIKE20 dataset.

Figure 12:

The ROC curve for each scenario on the PETS2006 dataset.
The ROC curve for each scenario on the PETS2006 dataset.

Figure 13:

The ROC curve for each scenario on the i-LIDS dataset.
The ROC curve for each scenario on the i-LIDS dataset.

The selected superpixels and their location based on bend line

Superpixels Location
B + 3h
B + 2h
B + 2h
B + h
B + h

Comparison of precision, recall, and F1 scores on the PETS2006 dataset

Methods Precision Recall F1 score
Damen and Hog (2012) 50% 55% 52%
Ghadiri et al. (2017) 57% 71% 63%
Ghadiri et al. (2019) 60% 79% 68%
Proposed Methods
BP_SC1 56% 85% 68%
BP_SC2 59% 83% 69%

SLIC segmentation

1: Centroid Initialization Ck=[lk,ak,bk,xk,yk]T
2: Put centroid in n × n window
3:   repeat
4: for each cluster Ck do
5: Group each pixel in the nearest centroid (based on measurement of pixel distance to centroid)
    end for
6: Update centroid
7: until centroid unchanged

Number of test images in each dataset

Dataset Test Images
DIKE20 271
PETS2006 323
i-LIDS 185
Total 779

The precision, recall, and F1 scores on DIKE20 dataset

Methods Precision Recall F1 score
BP_SC1 46% 79% 60%
BP_SC2 52% 80% 63%

Comparison of precision, recall, and F1 scores on the i-LIDS dataset

Methods Precision Recall F1 score
Damen and Hog (2012) 52% 47% 49%
Ghadiri et al. (2017) 62% 60% 61%
Ghadiri et al. (2019) 72% 64% 67%
Proposed Methods
BP_SC1 49% 90% 64%
BP_SC2 52% 95% 67%
eISSN:
1178-5608
Language:
English
Publication timeframe:
Volume Open
Journal Subjects:
Engineering, Introductions and Overviews, other
Journal RSS Feed