Recognition of partially occluded objects using neural network based indexing
Citations
39 citations
19 citations
16 citations
Cites background or methods from "Recognition of partially occluded o..."
...Even though there have been several efforts in object recognition with occlusion, currents methods have been highly sensitive to object pose, rotation, scaling, and visible portion of occluded objects [12] [9] [17] [3] [15]....
[...]
...[17] introduced a method for partial object recognition using neural network based indexing....
[...]
7 citations
Cites background from "Recognition of partially occluded o..."
...• Local invariants, on the other hand, such as geometric invariants (Rivlin and Weiss, 1995), dominant points (Rajpal, Chaudhury, and Banerjee, 1999) and boundary segments (Mehrotra and Gary, 1995), defined at each point of a shape seperately, can handle images containing partially visible,…...
[...]
7 citations
References
7,518 citations
"Recognition of partially occluded o..." refers background or methods in this paper
...Features extracted from these orientations have not been used in training phase of two types of neural networks and also not stored in the hash table....
[...]
...Number of mappings for which response was more than 0.5 after neural networks-based indexing, for both types of networks are given in the third and fourth row of the table....
[...]
...Similarity invariant feature vectors extracted from triplets of these images, were given as inputs to both types of neural networks and the output response corresponding to each model triplet was stored in the table....
[...]
...This 20-dimensional feature vector is used as input for neural network-based indexing technique explained in the next section....
[...]
2,411 citations
2,166 citations
"Recognition of partially occluded o..." refers background or methods in this paper
...Features extracted from these orientations have not been used in training phase of two types of neural networks and also not stored in the hash table....
[...]
...Number of mappings for which response was more than 0.5 after neural networks-based indexing, for both types of networks are given in the third and fourth row of the table....
[...]
...Similarity invariant feature vectors extracted from triplets of these images, were given as inputs to both types of neural networks and the output response corresponding to each model triplet was stored in the table....
[...]
...This 20-dimensional feature vector is used as input for neural network-based indexing technique explained in the next section....
[...]
1,395 citations
567 citations