Page 32 - Read Online
P. 32

assessment  has  been  an  active  area  of  research.   Veredas et al.  Finally, other authors have indicated that
                                                                          [15]
          Arnqvist  et  al.   addressed  segmenting  secondary  healing   an unaccounted  color  may appear in a wound bed and
                      [1]
          ulcers based on the 3‑color (RYK)  model: red (R)   affect classification. [17]
          granulation,  yellow (Y) necroses  and black (K) necrotic   The  current  study  proposes that  an  eight‑color  model
          eschar.  Herbin  et  al.,   in  their  study  on the  effect  of a   would  more  comprehensively  represent  wound
                            [2]
          new drug on wound kinetics,  used color information   composition, evolution and changes due to infection.
          (red  and green) to  distinguish  a  wound from  normal   This would assist in the development of a comprehensive
          skin  in  estimating  the  area  of  the  wound,  and proposed   algorithm for all wounds, rather than a specific wound
          a color index  to quantify  healing  on uniformly  colored,   type (e.g.  pressure ulcers). Thus, in the interest of
          artificially created blisters.  However, natural wounds are   improving  the accuracy of assessment  for general
          generally more complex with highly variegated coloring.    wounds, the authors discuss a knowledge base of clusters
                                                          [3]
          The use of color information, rather than size,  to assess   associated with the eight  categories in an appropriate
          healing and guide clinicians in their choice of wound care   color space. The results of a classification algorithm which
          products, was stressed by Mekkes and Westerhof.  The   take into account the information about the shape, spread
                                                     [4]
                                             [4]
          approach taken by Mekkes and Westerhof  in their study   and orientation  of each cluster,  through  its  covariance
          of wound debridement based on video images, uses the   matrix,  are presented.  In addition, the results of the
                                                                                  [18]
                                      [5]
          RYK model.  Jones  and Plassmann   used the  same  model   “rotated coordinate  system  (RCS)”  method,  which also
          along with an “unclassified” category in their description   makes use of cluster shapes, are delineated. [19]
          of skin ulcers. Hansen et al.  applied the hue, saturation
                                  [6]
          and intensity  (HSI) model to classify  wound severity  in
          an animal model within 30 min following injury. Wounds   MATERIALS AND METHODS
          were classified  as mild,  moderate or severe,  based on
                                                [7]
          differences  in  color.  Berris  and Sangwine   used  the   The procedure for computer assessment  of wound
          RGB model, in a study of pressure ulcers, to assess the   composition based on colors of tissue  and pigmentation
          content in  terms  of the  three  tissue  types  (BYR  model).   involves (a)  choosing an appropriate color  model,
                     [8]
          Hoppe  et  al.  highlighted the inadequacy of the RGB   (b) building the knowledge base in the chosen space, and
          model, and used the HSI model to grade leg ulcers in   (c) classification of regions in the given wound.
          terms of the quantity  of slough. They also investigated   The color model
          the variability of color attributes (only red was considered)   Although R, G and B are the considered the “primary colors”
          due to differences in lighting conditions and found hue   of light, the RGB model is more useful in the generation
                                            [9]
          to be  the least variant. Oduncu  et  al.  also used the   of color  rather than  analysis.  An alternative  is  the  HSI
                                                                                       [18]
          HSI model to assess  the  amount of slough in leg ulcers.   model, wherein hue (H) describes pure color, saturation (S)
          Varedas  et  al.   developed a  method  very  specific to   is  the  “degree  of dilution from  purity”  by  white  light,
                      [10]
          pressure  ulcers for classifying healthy  and nonhealthy   and I is the intensity (decoupled from color information).
          skin, with four tissue types for the latter, their method   Color in RGB‑format can be converted to the HSI version
          involves several preprocessing steps for the extraction of   by suitable transformations.  Since hue is quantified in
                                                                                      [18]
          color  and texture parameters. Wannous  et  al.  studied   terms  of the  angle  from  the  “red”‑axis  in  the  HSI  space,
                                                  [11]
          the  variability  arising  from  image  capture  from  different   it  suffers  discontinuity  at  0  (located at  the  mid‑point  of
          sources under different  conditions,  they  too considered   the interval representing the red hue). To circumvent this
          only three tissue types.   Dorileo  et  al.  added  white   problem, the hue range was shifted, as was suggested in
                                             [12]
          (W,  representing  hyperkeratosis) to the  RYK model, for   the previous study. [20]
          analyzing dermatological ulcers. In the same context, the
                                                                  360
                                                         [13]
          use of texture parameters was proposed by Pereira et al.    H  =   −  H  ∈ H,        [250, 360)   (1)
          to improve the performance of classification. The authors    H +  H  ∈110,        [0, 250)
          previously  presented  initial  results  showing  the  presence   The choice of the origin at H = 250 is guided by the fact
          of eight tissue types and pigmentation, with a one‑to‑one   that minimal tissue was found in the range of 250 ± 60.
          correspondence to distinct colors.  Subsequently,   We refer to this as the modified HSI space.
                                           [14]
          Veredas  et  al.,   in  their  elaborate  work  on  pressure
                       [15]
          ulcers based on four tissue‑types (with the skin    Procedure for assessing wound composition
          regarded as the fifth), have emphasized the necessity for   The algorithm for determining the wound status in terms
          precise evaluation. Recently, Mukherjee et al.  presented   of tissue types and pigmentation consists of two phases,
                                                [16]
          results using a three‑color model with textures.    as described in the following.
          Thus, while the importance of color  in representing   Building the knowledge base
          wound  composition is well‑recognized, only three to   The eight categories of tissue and pigmentation of
          four categories and colors have been  used to date   interest  include: (1) healthy  granulation tissue  (HGT),
          in the analysis of specific types of wounds. However,   (2) unhealthy granulation tissue (UGT), (3) whitish slough,
          wounds evolve over time,  with specific types of tissue   (4)  yellowish‑green pigmentation (G ), (5)  bluish‑green
                                                                                              1
          and pigmentation with  eight identifiable  colors which   pigmentation (G ), (6) fat (F), (7) brown necrotic tissue (BNT),
                                                                            2
          represent healing, necrosis and/or infection.  This theory   and (8)  black  necrotic tissue or gangrene (Ga).  A sample
                                               [14]
          is also supported by observations in a recent paper by   of each, pertaining to each of the eight categories, was
           262                                                           Plast Aesthet Res || Vol 2 || Issue 5 || Sep 15, 2015
   27   28   29   30   31   32   33   34   35   36   37