Integrating the Symmetry Image and Improved Sparse Representation for Railway Fastener Classification and Defect Recognition

Figure 4

A test sample that is erroneously and correctly classified by the Latent Dirichlet Allocation and our method, respectively. (a) The test sample. (b) The first symmetrical sample of the test sample. (c) The second symmetrical sample of test sample. (d) The training sample which is the nearest to the test sample. (e) The training sample which is the nearest to the first symmetrical sample. (f) The training sample which is the nearest to the second symmetrical sample.

(a)

(b)

(c)

(d)

(e)

(f)