Entropy optimized morphological shared-weight neural networks

Khabou, Mohamed A.; Gader, Paul D.; Shi, Hongchi

doi:10.1117/1.602085

Cited by 14 publications

(13 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, Won et al proposed a shared weight morphological network and applied it to target detection in [37]. Similar kinds of work can be found in [1] and [13].…”

Section: Introductionmentioning

confidence: 77%

Integrated feature analysis and fuzzy rule-based system identification in a neuro-fuzzy paradigm

Chakrborty

Pal

2001

IEEE Trans. Syst., Man, Cybern. B

View full text Add to dashboard Cite

Abstract-Most methods of fuzzy rule-based system identification (SI) either ignore feature analysis or do it in a separate phase. This paper proposes a novel neuro-fuzzy system that can simultaneously do feature analysis and SI in an integrated manner. It is a five-layered feed-forward network for realizing a fuzzy rule-based system. The second layer of the net is the most important one, which along with fuzzification of the input also learns a modulator function for each input feature. This enables online selection of important features by the network. The system is so designed that learning maintains the nonnegative characteristic of certainty factors of rules. The proposed network is tested on both synthetic and real data sets and the performance is found to be quite satisfactory. To get an "optimal" network architecture and to eliminate conflicting rules, nodes and links are pruned and then the structure is retrained. The pruned network retains almost the same level of performance as that of the original one.

show abstract

“…For example, Won et al proposed a shared weight morphological network and applied it to target detection in [37]. Similar kinds of work can be found in [1] and [13].…”

Section: Introductionmentioning

confidence: 77%

Integrated feature analysis and fuzzy rule-based system identification in a neuro-fuzzy paradigm

Chakrborty

Pal

2001

IEEE Trans. Syst., Man, Cybern. B

View full text Add to dashboard Cite

show abstract

“…Mathematical morphology is a useful technique for image processing, pattern recognition, shape analysis, and feature extraction [15]. Morphological shared-weight neural networks (MSNNs) have been successfully used for ATR research from multiple sensors including infrared, optical camera, SAR, and LADAR [16][17][18][19]. The morphological structure elements are learned as feature extractors simultaneously with the classification parameters.…”

Section: Introductionmentioning

confidence: 99%

Vehicle detection from high-resolution satellite imagery using morphological shared-weight neural networks

Jin

Davis

2007

Image and Vision Computing

View full text Add to dashboard Cite

“…The Blazer is not occluded in any of the training images and the level of occlusion in the testing images ranges from no occlusion to complete occlusion. A detection rate of 39/52 with 33 false alarms was reported in [17]. It is important to note that neither the MACE nor the SSNN with a structure identical to that of the MSNN were able to achieve such results.…”

Section: Msnn Atr Results With Visible Spectrum Imagerymentioning

confidence: 86%

“…[14,15] and Khabou et al [17] have used the MSNN to detect a particular type of vehicle, a Chevrolet Blazer, in gray-scale visible images of a parking lot filled with other types of vehicles. The data set used was the Blazer data set collected at Eglin Air Force Base (Fig.…”

Section: Msnn Atr Results With Visible Spectrum Imagerymentioning

confidence: 99%

See 1 more Smart Citation

Morphological shared-weight neural networks: a tool for automatic target recognition beyond the visible spectrum

Khabou

Gader²,

Keller³

Proceedings IEEE Workshop on Computer Vision Beyond the Visible Spectrum: Methods and Applications (CVBVS'99)

View full text Add to dashboard Cite

Morphological shared-weight neural networks (MSNN) combine the feature extraction capability of mathematical morphology with the function mapping capability of neural networks. This provides a trainable mechanism for translation invariant object detection using a variety of imaging sensors, including TV, forward-looking infrared (FLIR) and synthetic aperture radar (SAR). In this paper, we provide an overview of previous results and new results with laser radar (LADAR). We present three sets of experiments. In the first set of experiments we use the MSNN to detect different types of targets simultaneously. In the second set we use the MSNN to detect only a particular type of target. In the third set we test a novel scenario: We train the MSNN to recognize a particular type of target using very few examples. A detection rate of 86% with a reasonable number of false alarms was achieved in the first set of experiments and a detection rate of close to 100% with very few false alarms was achieved in the second and third sets of experiments.

show abstract

Entropy optimized morphological shared-weight neural networks

Cited by 14 publications

References 7 publications

Integrated feature analysis and fuzzy rule-based system identification in a neuro-fuzzy paradigm

Integrated feature analysis and fuzzy rule-based system identification in a neuro-fuzzy paradigm

Vehicle detection from high-resolution satellite imagery using morphological shared-weight neural networks

Morphological shared-weight neural networks: a tool for automatic target recognition beyond the visible spectrum

Contact Info

Product

Resources

About