Figure 1.
The general pipeline of a typical methodology which utilizes a spatial neighborhood method technique.
Figure 1.
The general pipeline of a typical methodology which utilizes a spatial neighborhood method technique.
Figure 2.
The pattern emitted by the projector.
Figure 2.
The pattern emitted by the projector.
Figure 3.
The 6 symbols used to create the pattern.
Figure 3.
The 6 symbols used to create the pattern.
Figure 4.
The sequence of symbol recording for transforming a subpattern into a codeword and vice versa, s1–s9 represent the symbols of a subpattern.
Figure 4.
The sequence of symbol recording for transforming a subpattern into a codeword and vice versa, s1–s9 represent the symbols of a subpattern.
Figure 5.
(a) The RGB image captured by the camera, (b) the corresponding grayscale image, (c) the resulting binary image after Sauvola’s method.
Figure 5.
(a) The RGB image captured by the camera, (b) the corresponding grayscale image, (c) the resulting binary image after Sauvola’s method.
Figure 6.
The successive steps for extracting the nearest symbol for each symbol of the subpattern. The symbol we visit in the current iteration is marked in red, and the symbol that represents its closest neighboring symbol within the subpattern is marked in green.
Figure 6.
The successive steps for extracting the nearest symbol for each symbol of the subpattern. The symbol we visit in the current iteration is marked in red, and the symbol that represents its closest neighboring symbol within the subpattern is marked in green.
Figure 7.
The four grids of neighboring symbols which compose the subpattern. The four symbols of each grid are marked by green dots.
Figure 7.
The four grids of neighboring symbols which compose the subpattern. The four symbols of each grid are marked by green dots.
Figure 8.
The successive steps for extracting the four grids of symbols. In (a), the symbols that form the first grid are marked in red, while in (b–d), symbols extracted for the first time in the second, third, and fourth steps of the corresponding algorithm, respectively, are marked in green, blue, and cyan color.
Figure 8.
The successive steps for extracting the four grids of symbols. In (a), the symbols that form the first grid are marked in red, while in (b–d), symbols extracted for the first time in the second, third, and fourth steps of the corresponding algorithm, respectively, are marked in green, blue, and cyan color.
Figure 9.
The successive steps for projecting a group of 4 points onto their BB vertices, with green dots representing the 4 points and white stars indicating the BB vertices that are not considered in subsequent algorithm steps.
Figure 9.
The successive steps for projecting a group of 4 points onto their BB vertices, with green dots representing the 4 points and white stars indicating the BB vertices that are not considered in subsequent algorithm steps.
Figure 10.
Two separate grids of symbols with a common symbol, specifically a triangle, marked with a red dot. The remaining 3 symbols of each grid are marked with green dots. In the left grid, the triangle is the bottom-left symbol, while in the right grid, the triangle is the top-left symbol.
Figure 10.
Two separate grids of symbols with a common symbol, specifically a triangle, marked with a red dot. The remaining 3 symbols of each grid are marked with green dots. In the left grid, the triangle is the bottom-left symbol, while in the right grid, the triangle is the top-left symbol.
Figure 11.
Four grids of symbols with a common symbol, specifically a ‘Π’ symbol, marked with a red dot. The remaining 3 symbols of each grid are marked with green dots. In the various grids, this symbol occupies all four possible positions it can be in, specifically: bottom-left in (a), bottom-right in (b), top-right in (c), and top-left in (d).
Figure 11.
Four grids of symbols with a common symbol, specifically a ‘Π’ symbol, marked with a red dot. The remaining 3 symbols of each grid are marked with green dots. In the various grids, this symbol occupies all four possible positions it can be in, specifically: bottom-left in (a), bottom-right in (b), top-right in (c), and top-left in (d).
Figure 12.
The projector–camera system. The distance between them, the respective Fields Of View, and the distance at which the fishes are positioned are also indicated.
Figure 12.
The projector–camera system. The distance between them, the respective Fields Of View, and the distance at which the fishes are positioned are also indicated.
Figure 13.
The fishes used for the experiments: (a) white seabream, (b) red seabream, (c) gilthead seabream, (d) European seabass.
Figure 13.
The fishes used for the experiments: (a) white seabream, (b) red seabream, (c) gilthead seabream, (d) European seabass.
Figure 14.
The images captured by the camera after the pattern falls onto the bodies of the fishes: (a) white seabream, (b) red seabream, (c) gilthead seabream, (d) European seabass.
Figure 14.
The images captured by the camera after the pattern falls onto the bodies of the fishes: (a) white seabream, (b) red seabream, (c) gilthead seabream, (d) European seabass.
Figure 15.
The resulting binarized images: (a) white seabream, (b) red seabream, (c) gilthead seabream, (d) European seabass.
Figure 15.
The resulting binarized images: (a) white seabream, (b) red seabream, (c) gilthead seabream, (d) European seabass.
Figure 16.
The histograms of the minimum Hamming distances between the codewords extracted from the images and the valid codewords of the pattern, as a function of the possible corresponding minimum Hamming distances: (a) white seabream, (b) red seabream, (c) gilthead seabream, (d) European seabass.
Figure 16.
The histograms of the minimum Hamming distances between the codewords extracted from the images and the valid codewords of the pattern, as a function of the possible corresponding minimum Hamming distances: (a) white seabream, (b) red seabream, (c) gilthead seabream, (d) European seabass.
Figure 17.
The final 3D reconstruction of (a) white seabream, (b) red seabream, (c) gilthead seabream, (d) European seabass.
Figure 17.
The final 3D reconstruction of (a) white seabream, (b) red seabream, (c) gilthead seabream, (d) European seabass.
Figure 18.
The depth maps corresponding to the fishes reconstructed in our paper: (a) white seabream, (b) red seabream, (c) gilthead seabream, (d) European seabass. On the right, a color bar is displayed, representing the color-depth mapping for various points in the 3D reconstructions.
Figure 18.
The depth maps corresponding to the fishes reconstructed in our paper: (a) white seabream, (b) red seabream, (c) gilthead seabream, (d) European seabass. On the right, a color bar is displayed, representing the color-depth mapping for various points in the 3D reconstructions.
Figure 19.
The Ground Truth (GT) models of the fishes reconstructed in our paper: (a) white seabream, (b) red seabream, (c) gilthead seabream, (d) European seabass, and the corresponding signed distances histogram between the points resulting from our 3D reconstructions and their corresponding nearest points in the reconstructions derived from the GT data. The colored points on the GT models represent the points resulting from our 3D reconstructions. The colors of these points correspond to the colors of the signed distances shown in the histograms above.
Figure 19.
The Ground Truth (GT) models of the fishes reconstructed in our paper: (a) white seabream, (b) red seabream, (c) gilthead seabream, (d) European seabass, and the corresponding signed distances histogram between the points resulting from our 3D reconstructions and their corresponding nearest points in the reconstructions derived from the GT data. The colored points on the GT models represent the points resulting from our 3D reconstructions. The colors of these points correspond to the colors of the signed distances shown in the histograms above.
Figure 20.
Binarization results using Otsu’s method for (a) white seabream, (b) red seabream, (c) gilthead seabream, (d) European seabass.
Figure 20.
Binarization results using Otsu’s method for (a) white seabream, (b) red seabream, (c) gilthead seabream, (d) European seabass.
Figure 21.
The number of codewords in which each symbol participates. The different colors indicate the corresponding number of each codeword, according to the column at the right of the image.
Figure 21.
The number of codewords in which each symbol participates. The different colors indicate the corresponding number of each codeword, according to the column at the right of the image.
Table 1.
SNR (in dB) of the images captured by the camera for each of the four fish cases used in the experiments.
Table 1.
SNR (in dB) of the images captured by the camera for each of the four fish cases used in the experiments.
Fish Type | SNR (in dB) |
---|
white seabream | 13.68 |
red seabream | 15.90 |
gilthead seabream | 14.62 |
European seabass | 13.10 |
Table 2.
The number of symbols contained in the binarized images, as well as the number of valid codewords contained in these images and the final number of corresponding points for the case of white seabream.
Table 2.
The number of symbols contained in the binarized images, as well as the number of valid codewords contained in these images and the final number of corresponding points for the case of white seabream.
Number of symbols in the image, | 380 |
Number of valid codewords | 248 |
Final number of corresponding points between the pattern and the image, | 362 |
Table 3.
The number of symbols contained in the binarized images, as well as the number of valid codewords contained in these images and the final number of corresponding points for the case of red seabream.
Table 3.
The number of symbols contained in the binarized images, as well as the number of valid codewords contained in these images and the final number of corresponding points for the case of red seabream.
Number of symbols in the image, | 275 |
Number of valid codewords | 165 |
Final number of corresponding points between the pattern and the image, | 254 |
Table 4.
The number of symbols contained in the binarized images, as well as the number of valid codewords contained in these images and the final number of corresponding points for the case of gilthead seabream.
Table 4.
The number of symbols contained in the binarized images, as well as the number of valid codewords contained in these images and the final number of corresponding points for the case of gilthead seabream.
Number of symbols in the image, | 437 |
Number of valid codewords | 260 |
Final number of corresponding points between the pattern and the image, | 387 |
Table 5.
The number of symbols contained in the binarized images, as well as the number of valid codewords contained in these images and the final number of corresponding points for the case of European seabass.
Table 5.
The number of symbols contained in the binarized images, as well as the number of valid codewords contained in these images and the final number of corresponding points for the case of European seabass.
Number of symbols in the image, | 330 |
Number of valid codewords | 189 |
Final number of corresponding points between the pattern and the image, | 320 |
Table 6.
The normalized SAD matrix, which contains the normalized SAD of each symbol from each other.
Table 6.
The normalized SAD matrix, which contains the normalized SAD of each symbol from each other.
| Symbol | Rectangle | ‘L’ | ‘T’ | Triangle | ‘X’ | ‘Π’ |
---|
Symbol | |
---|
Rectangle | 0 | 1.31 | 1.31 | 0.97 | 1.11 | 1.17 |
‘L’ | 1.31 | 0 | 1.50 | 0.89 | 1.17 | 0.93 |
‘T’ | 1.31 | 1.50 | 0 | 0.98 | 1.03 | 1.06 |
Triangle | 0.97 | 0.89 | 0.98 | 0 | 0.73 | 1.47 |
‘X’ | 1.11 | 1.17 | 1.03 | 0.73 | 0 | 1.29 |
‘Π’ | 1.17 | 0.93 | 1.06 | 1.47 | 1.29 | 0 |
Table 7.
The actual total length and height as well as the experimentally evaluated measurements obtained for the four fishes used in our experiments. All measurements are in cm.
Table 7.
The actual total length and height as well as the experimentally evaluated measurements obtained for the four fishes used in our experiments. All measurements are in cm.
| Fish | White Seabream | Red Seabream | Gilthead Seabream | European Seabass |
---|
Measurement | |
---|
Total length resulting from 3D reconstruction | 29.8 | 27.5 | 32.7 | 31.2 |
Actual total length | 30.6 | 26.0 | 31.7 | 31.7 |
Height resulting from 3D reconstruction | 11.2 | 11.0 | 10.9 | 7.4 |
Actual height | 11.2 | 10.4 | 11.0 | 7.8 |