Revisited histogram equalization as HDR images tone mapping operators

Abstract

High Dynamic Range (HDR) image provides higher perceptual quality such that it appears considerably more realistic and attractive for the human observer. Since most of current screens are Low Dynamic Range (LDR) screens, lots of researches have been proposed to design tone mapping algorithms converting the HDR images into a range that is suitable to display these tone mapped images on standard LDR screens. For this purpose, this paper first investigates the pixels distribution in the HDR image through the study of various traditional Histogram Equalization (HE) algorithms which were originally developed to improve the visual quality of LDR images such as the contrast enhancement of the latter. Moreover, a modification of the Histogram Adjustment based Linear to Equalized Quantizer (HALEQ), developed for HDR images, is proposed. Simulation results show that the proposed modification preserves more details than the original version of the algorithm in most parts of the HDR image.

Performance Analysis of the Revisited Tone Mapped Quality Index for Tone Mapped HDR Images Evaluation

“…P LCC ranges from −1 and +1. Fifteen TMOs are used for the training such as "Drago" [7], "Reinhard" [8], "Ward" [9], "Durand" [10], "Tumblin" [11], "Schlick" [12], "Duan" [13], "Fattal" WRB [14], "Li" [15], "Husseis" [16], "NUHA" [19], "SEPENO" [17], "NONSEPENO" [18], "CEDP Lin" [20] with β = 0.25 and "CEDP Opt" [20] with adaptive β. The training TM HDR image database is based on 24 HDR images with different dynamic range (or contrast ratio) from 7 f-stops to 29 f-stops, namely "Lausanne1", "CraterLake1", "Shasta2", "Synagogue", "Anturium", "BowRiver", "Bridges", "Stairway1", "ArchRock", "DollDoll", "ClockBuilding", "OxfordChurch", "Bot-tlesSmall", "Montreal", "SmallOffice", "Light", "BridgeStudios2", "Memorial", "ClaridgeHotel", "Mistaya1", "BrookHouse", "Peace-Rocks", "GGpark2" and "AtriumNight".…”

“…Indeed this metric does not take into account the new distortions introduced by the new TMOs (e.g. "Duan" [13], "Fattal" WRB [14], "Li" [15], "Husseis" [16], "NUHA" [19], "SEPENO" [17], "NON-SEPENO" [18], "CEDP" [20]). That is why this paper addresses, in section III-C, the question of re-evaluating the TMQI parameters to be in accordance with the MOS.…”

Tone Mapped HDR Images Contrast Enhancement Using Piecewise Linear Perceptual Transformation

“…where β i is a positive parameter smaller than 1 depending on the sub-interval [c i uHDR , c i nuHDR ]. Since the lower cutting point l i HDR (1) is deduced as the average of the logarithm luminance values on each sub-interval [c i uHDR , c i nuHDR ] as in [13], the parameter β i is then given by:…”

“…In [12], a TM optimization approach using a histogram adjustment between linear mapping and the equalized histogram mapping is developed. A modification of this approach is made in [13]. The latter considers both histogram equalization and human sensitivity to the light function.…”

HDR Image Tone Mapping Approach based on Near Optimal Separable Adaptive Lifting Scheme

“…In [9], a TM optimization approach using a histogram adjustment between linear mapping and the equalized histogram mapping is developed. A modification of this approach is made where revisited histogram equalization approaches are discussed in [10]. The latter considers both histogram equalization and human sensitivity to the light function.…”