A Study of Feature Combination for Vehicle Detection Based on Image Processing

Table 13

Performance of the proposed ensembles for the far region.

(a)

PCA + V-HOG; + LG+𝑆+𝐺	Accuracy	Recall	Precision
RMM-SA	97.16	97.12	97.20
RMM-WA	98.34	98.24	98.44
DS-SA	96.98	96.85	97.12
DS-WA	98.16	98.28	98.04

(b)

PCA+V-HOG; +LG+𝑆	Accuracy	Recall	Precision
RMM-SA	98.38	97.12	95.60
RMM-WA	98.14	98.35	97.92
DS-SA	96.14	96.88	95.36
DS-WA	98.18	98.39	97.96

(c)

PCA+V-HOG + 𝑆	Accuracy	Recall	Precision
RMM-SA	94.52	96.81	92.08
RMM-WA	97.34	98.85	95.80
DS-SA	94.24	96.67	91.64
DS-WA	97.42	98.85	95.96

(d)

PCA + LG+𝑆	Accuracy	Recall	Precision
RMM-SA	94.52	94.67	94.36
RMM-WA	96.76	95.95	97.64
DS-SA	94.22	94.42	94.00
DS-WA	96.76	95.92	97.68

(e)

PCA+V-HOG+LG	Accuracy	Recall	Precision
RMM-SA	97.22	98.05	96.36
RMM-WA	97.26	97.97	96.52
DS-SA	97.24	98.09	96.36
DS-WA	97.24	98.01	96.44