Modified pose adaption so that it can now refer face from another pose, Added node to compare body proportion ratio for the given 2 poses and body part

__init__.py

@@ -1155,20 +1155,25 @@ class TRI3DPoseAdaption:
                     "max": 15.0,
                     "step": 0.01
                 }),
-                "garment_category":([" no_sleeve_garment", "half_sleeve_garment", "full_sleeve_garment", \
-                                     "shorts", "trouser"], {"default": " no_sleeve_garment"})
-
-
+                "garment_category":(["no_sleeve_garment", "half_sleeve_garment", "full_sleeve_garment", \
+                                     "shorts", "trouser"], {"default": "no_sleeve_garment"}),
+                "output_save_path": ("STRING",{"default" : "dwpose/keypoints/adapted_pose.json"})
+            },
+            
+            "optional": {
+                "ref_face_pose_json_file": ("STRING",{"default" : ""}),
             }
         }
+        
 
     RETURN_TYPES = ("IMAGE", "BOOLEAN")
     FUNCTION = "main"
     CATEGORY = "TRI3D"
 
-    def main(self, input_pose_json_file, ref_pose_json_file, image_angle, rotation_threshold, garment_category):
+    def main(self, input_pose_json_file, ref_pose_json_file, image_angle, rotation_threshold, garment_category,output_save_path, ref_face_pose_json_file=""):
         from .dwpose import comfy_utils
-
+        cur_file_dir = os.path.dirname(os.path.realpath(__file__))
+        output_save_path = os.path.join(cur_file_dir, output_save_path)
         if image_angle == "front":
 
             input_pose = json.load(open(input_pose_json_file))
@@ -1181,6 +1186,13 @@ class TRI3DPoseAdaption:
             ref_height = ref_pose['height']
             ref_width = ref_pose['width']
             ref_keypoints = ref_pose['keypoints']
+
+            if ref_face_pose_json_file != "":
+                ref_face_pose = json.load(open(ref_face_pose_json_file))
+                ref_face_height = ref_face_pose['height']
+                ref_face_width = ref_face_pose['width']
+                ref_face_keypoints = ref_face_pose['keypoints']
+                
             similar_torso = None
 
             #check torso similarity
@@ -1204,6 +1216,11 @@ class TRI3DPoseAdaption:
                     ]
                     return (canvas, similar_torso)
             
+
+            #Scaling shoulder of input with respect to shoulder-torso ratio of ref-pose
+            input_keypoints = comfy_utils.scale(ref_keypoints, input_keypoints, 1, 2, 1, 2, ref_torso=True)             #left shoulder
+            input_keypoints = comfy_utils.scale(ref_keypoints, input_keypoints, 1, 5, 1, 5, ref_torso=True)             #right shoulder
+
             #Hands
             if input_keypoints[4] == [-1,-1]: input_keypoints[4] = ref_keypoints[4]
             if input_keypoints[7] == [-1,-1]: input_keypoints[7] = ref_keypoints[7]
@@ -1242,11 +1259,10 @@ class TRI3DPoseAdaption:
             # input_keypoints = comfy_utils.rotate_hand(ref_keypoints, input_keypoints, 88)    #rotating right hand
             input_keypoints = comfy_utils.scale_hand(ref_keypoints, input_keypoints, 6, 7, 88)    #scaling right hand w.r.t right wrist
 
-
             #legs
             if garment_category not in ["trouser", "shorts"]:
                 input_keypoints = comfy_utils.rotate(ref_keypoints, input_keypoints, 8, 9)      #rotate left knee
-            input_keypoints = comfy_utils.scale(ref_keypoints, input_keypoints, 1, 8, 8, 9)      #scale left knee
+            input_keypoints = comfy_utils.scale(ref_keypoints, input_keypoints, 1, 8, 8, 9, ref_torso=True)      #scale left knee
             
             if garment_category != "trouser":
                 input_keypoints = comfy_utils.rotate(ref_keypoints, input_keypoints, 9, 10)     #rotate left foot
@@ -1254,7 +1270,7 @@ class TRI3DPoseAdaption:
             
             if garment_category not in ["trouser", "shorts"]:
                 input_keypoints = comfy_utils.rotate(ref_keypoints, input_keypoints, 11, 12)     #rotate right knee
-            input_keypoints = comfy_utils.scale(ref_keypoints, input_keypoints, 1, 11, 11, 12)      #scale right knee
+            input_keypoints = comfy_utils.scale(ref_keypoints, input_keypoints, 1, 11, 11, 12, ref_torso=True)      #scale right knee
             
             if garment_category != "trouser":
                 input_keypoints = comfy_utils.rotate(ref_keypoints, input_keypoints, 12, 13)    #rotate right foot
@@ -1262,40 +1278,48 @@ class TRI3DPoseAdaption:
 
             #face
             prev_nose = input_keypoints[0]
-            input_keypoints = comfy_utils.rotate(ref_keypoints, input_keypoints, 1, 0)      #rotate nose
-            input_keypoints = comfy_utils.scale(ref_keypoints, input_keypoints, 2, 5, 1, 0)      #scale neck to nose w.r.t to shoulder neck-nose len ratio of ref pose
-            
+
+            if ref_face_pose_json_file != "":
+                reference = ref_face_keypoints
+            else:
+                reference = ref_keypoints
+
+            input_keypoints = comfy_utils.rotate(reference, input_keypoints, 1, 0)      #rotate nose
+            input_keypoints = comfy_utils.scale(reference, input_keypoints, 2, 5, 1, 0)      #scale neck to nose w.r.t to shoulder neck-nose len ratio of ref pose
+                
             #changing face points to w.r.t to new nose point after rotation
             input_keypoints[14:18] = comfy_utils.move(prev_nose, input_keypoints[0], input_keypoints[14:18])
 
-            input_keypoints = comfy_utils.rotate(ref_keypoints, input_keypoints, 0,
+            input_keypoints = comfy_utils.rotate(reference, input_keypoints, 0,
                                                 14)  #rotate left eye
             input_keypoints = comfy_utils.scale(
-                ref_keypoints, input_keypoints, 1, 0, 0,
+                reference, input_keypoints, 1, 0, 0,
                 14)  #scaling w.r.t to neck len to eye_nose len ratio of ref pose
 
-            input_keypoints = comfy_utils.rotate(ref_keypoints, input_keypoints, 0,
+            input_keypoints = comfy_utils.rotate(reference, input_keypoints, 0,
                                                 15)  #rotate right eye
             input_keypoints = comfy_utils.scale(
-                ref_keypoints, input_keypoints, 1, 0, 0,
+                reference, input_keypoints, 1, 0, 0,
                 15)  #scaling w.r.t to neck len to eye_nose len ratio of ref pose
 
-            input_keypoints = comfy_utils.rotate(ref_keypoints, input_keypoints,
+            input_keypoints = comfy_utils.rotate(reference, input_keypoints,
                                                 14, 16)  #rotate left ear
             input_keypoints = comfy_utils.scale(
-                ref_keypoints, input_keypoints, 1, 0, 14,
+                reference, input_keypoints, 1, 0, 14,
                 16)  #scaling w.r.t to neck len to ear_nose len ratio of ref pose
 
-            input_keypoints = comfy_utils.rotate(ref_keypoints, input_keypoints,
+            input_keypoints = comfy_utils.rotate(reference, input_keypoints,
                                                 15, 17)  #rotate right ear
             input_keypoints = comfy_utils.scale(
-                ref_keypoints, input_keypoints, 1, 0, 15,
+                reference, input_keypoints, 1, 0, 15,
                 17)  #scaling w.r.t to neck len to ear_nose len ratio of ref pose
 
             canvas = comfy_utils.draw_bodypose(canvas, input_keypoints)
             canvas = comfy_utils.draw_handpose(canvas, input_keypoints[88:109])  #right hand
             canvas = comfy_utils.draw_handpose(canvas, input_keypoints[109:])  #left hand
             canvas = torch.from_numpy(canvas.astype(np.float32)/255.0)[None,]
+            output_posemap = {"height":input_height, "width":input_width, "keypoints":input_keypoints}
+            json.dump(output_posemap, open(output_save_path, 'w'))
             return (canvas, similar_torso)
         
         if 'back' in image_angle:
@@ -1452,6 +1476,8 @@ class TRI3DPoseAdaption:
             canvas = comfy_utils.draw_handpose(canvas, input_keypoints[88:109])  #right hand
             canvas = comfy_utils.draw_handpose(canvas, input_keypoints[109:])  #left hand
             canvas = torch.from_numpy(canvas.astype(np.float32)/255.0)[None,]
+            output_posemap = {"height":input_height, "width":input_width, "keypoints":input_keypoints}
+            json.dump(output_posemap, open(output_save_path, 'w'))
             return (canvas, similar_torso)
 
 
@@ -2086,6 +2112,100 @@ class TRI3D_clipdrop_bgremove_api:
         # print(mask.shape)
         return output,
 
+class TRI3DPoseProportion:
+    
+    def __init__(self):
+        pass
+
+    @classmethod
+    def INPUT_TYPES(s):
+        return {"required":
+                    {"pose1_json_file":("STRING",{"default":""}), "pose2_json_file":("STRING",{"default":""}), \
+                     "body_part":(["face", "left_elbow", "right_elbow", "left_wrist", "right_wrist", \
+                                   "left_knee", "right_knee", "left_ankle", "right_ankle"],{"default":"face"})
+                     }
+                }
+
+    FUNCTION = "run"
+    RETURN_TYPES = ("FLOAT","FLOAT")
+    RETURN_NAMES = ("pose1_proportion", "pose2_proportion")
+    CATEGORY = "TRI3D"
+
+    def run(self, pose1_json_file, pose2_json_file, body_part):
+        from .dwpose import comfy_utils
+
+        pose1 = json.load(open(pose1_json_file,'r'))
+        pose2 = json.load(open(pose2_json_file,'r'))
+        
+        pose1_keypoints = pose1['keypoints']
+        pose2_keypoints = pose2['keypoints']
+
+        #Calculate torso 2nd coord
+        pose1_ref_torso_x = int((pose1_keypoints[8][0] + pose1_keypoints[11][0]) / 2)
+        pose1_ref_torso_y = int((pose1_keypoints[8][1] + pose1_keypoints[11][1]) / 2)
+        pose2_ref_torso_x = int((pose2_keypoints[8][0] + pose2_keypoints[11][0]) / 2)
+        pose2_ref_torso_y = int((pose2_keypoints[8][1] + pose2_keypoints[11][1]) / 2)
+
+        if body_part == "face":
+            pose1_ratio = comfy_utils.get_ratio(pose1_keypoints[2][0], pose1_keypoints[2][1], pose1_keypoints[5][0], pose1_keypoints[5][1],                       
+                            pose1_keypoints[16][0], pose1_keypoints[16][1], pose1_keypoints[17][0], pose1_keypoints[17][1])
+
+            pose2_ratio = comfy_utils.get_ratio(pose2_keypoints[2][0], pose2_keypoints[2][1], pose2_keypoints[5][0], pose2_keypoints[5][1],                       
+                            pose2_keypoints[16][0], pose2_keypoints[16][1], pose2_keypoints[17][0], pose2_keypoints[17][1])
+        elif body_part == "left_elbow":
+            pose1_ratio = comfy_utils.get_ratio(pose1_keypoints[2][0], pose1_keypoints[2][1], pose1_keypoints[5][0], pose1_keypoints[5][1],                       
+                            pose1_keypoints[2][0], pose1_keypoints[2][1], pose1_keypoints[3][0], pose1_keypoints[3][1])
+
+            pose2_ratio = comfy_utils.get_ratio(pose2_keypoints[2][0], pose2_keypoints[2][1], pose2_keypoints[5][0], pose2_keypoints[5][1],                       
+                            pose2_keypoints[2][0], pose2_keypoints[2][1], pose2_keypoints[3][0], pose2_keypoints[3][1])
+        elif body_part == "right_elbow":
+            pose1_ratio = comfy_utils.get_ratio(pose1_keypoints[2][0], pose1_keypoints[2][1], pose1_keypoints[5][0], pose1_keypoints[5][1],                       
+                            pose1_keypoints[5][0], pose1_keypoints[5][1], pose1_keypoints[6][0], pose1_keypoints[6][1])
+
+            pose2_ratio = comfy_utils.get_ratio(pose2_keypoints[2][0], pose2_keypoints[2][1], pose2_keypoints[5][0], pose2_keypoints[5][1],                       
+                            pose2_keypoints[5][0], pose2_keypoints[5][1], pose2_keypoints[6][0], pose2_keypoints[6][1])
+        elif body_part == "left_wrist":
+            pose1_ratio = comfy_utils.get_ratio(pose1_keypoints[2][0], pose1_keypoints[2][1], pose1_keypoints[3][0], pose1_keypoints[3][1],                       
+                            pose1_keypoints[3][0], pose1_keypoints[3][1], pose1_keypoints[4][0], pose1_keypoints[4][1])
+
+            pose2_ratio = comfy_utils.get_ratio(pose2_keypoints[2][0], pose2_keypoints[2][1], pose2_keypoints[3][0], pose2_keypoints[3][1],                       
+                            pose2_keypoints[3][0], pose2_keypoints[3][1], pose2_keypoints[4][0], pose2_keypoints[4][1])
+        elif body_part == "right_wrist":
+            pose1_ratio = comfy_utils.get_ratio(pose1_keypoints[5][0], pose1_keypoints[5][1], pose1_keypoints[6][0], pose1_keypoints[6][1],                       
+                            pose1_keypoints[6][0], pose1_keypoints[6][1], pose1_keypoints[7][0], pose1_keypoints[7][1])
+
+            pose2_ratio = comfy_utils.get_ratio(pose2_keypoints[5][0], pose2_keypoints[5][1], pose2_keypoints[6][0], pose2_keypoints[6][1],                       
+                            pose2_keypoints[6][0], pose2_keypoints[6][1], pose2_keypoints[7][0], pose2_keypoints[7][1])
+        elif body_part == "left_knee":
+            pose1_ratio = comfy_utils.get_ratio(pose1_keypoints[1][0], pose1_keypoints[1][1], pose1_ref_torso_x, pose1_ref_torso_y,                       
+                            pose1_keypoints[8][0], pose1_keypoints[8][1], pose1_keypoints[9][0], pose1_keypoints[9][1])
+
+            pose2_ratio = comfy_utils.get_ratio(pose2_keypoints[1][0], pose2_keypoints[1][1], pose2_ref_torso_x, pose2_ref_torso_y,                       
+                            pose2_keypoints[8][0], pose2_keypoints[8][1], pose2_keypoints[9][0], pose2_keypoints[9][1])
+
+        elif body_part == "right_knee":
+            pose1_ratio = comfy_utils.get_ratio(pose1_keypoints[1][0], pose1_keypoints[1][1], pose1_ref_torso_x, pose1_ref_torso_y,                       
+                            pose1_keypoints[11][0], pose1_keypoints[11][1], pose1_keypoints[12][0], pose1_keypoints[12][1])
+
+            pose2_ratio = comfy_utils.get_ratio(pose2_keypoints[1][0], pose2_keypoints[1][1], pose2_ref_torso_x, pose2_ref_torso_y,                       
+                            pose2_keypoints[11][0], pose2_keypoints[11][1], pose2_keypoints[12][0], pose2_keypoints[12][1])
+        
+        elif body_part == "left_ankle":
+            pose1_ratio = comfy_utils.get_ratio(pose1_keypoints[8][0], pose1_keypoints[8][1], pose1_keypoints[9][0], pose1_keypoints[9][1],                       
+                            pose1_keypoints[9][0], pose1_keypoints[9][1], pose1_keypoints[10][0], pose1_keypoints[10][1])
+
+            pose2_ratio = comfy_utils.get_ratio(pose2_keypoints[8][0], pose2_keypoints[8][1], pose2_keypoints[9][0], pose2_keypoints[9][1],                       
+                            pose2_keypoints[9][0], pose2_keypoints[9][1], pose2_keypoints[10][0], pose2_keypoints[10][1])
+
+        elif body_part == "right_ankle":
+            pose1_ratio = comfy_utils.get_ratio(pose1_keypoints[11][0], pose1_keypoints[11][1], pose1_keypoints[12][0], pose1_keypoints[12][1],                       
+                            pose1_keypoints[12][0], pose1_keypoints[12][1], pose1_keypoints[13][0], pose1_keypoints[13][1])
+
+            pose2_ratio = comfy_utils.get_ratio(pose2_keypoints[11][0], pose2_keypoints[11][1], pose2_keypoints[12][0], pose2_keypoints[12][1],                       
+                            pose2_keypoints[12][0], pose2_keypoints[12][1], pose2_keypoints[13][0], pose2_keypoints[13][1])
+        
+        return pose1_ratio, pose2_ratio
+
 # A dictionary that contains all nodes you want to export with their names
 # NOTE: names should be globally unique
 NODE_CLASS_MAPPINGS = {
@@ -2107,7 +2227,8 @@ NODE_CLASS_MAPPINGS = {
     "tri3d-recolor-mask-RGB_space": TRI3D_recolor_RGB,
     "tri3d-image-mask-2-box": TRI3D_image_mask_2_box,
     "tri3d-image-mask-box-2-image": TRI3D_image_mask_box_2_image,
-    "tri3d-clipdrop-bgremove-api": TRI3D_clipdrop_bgremove_api
+    "tri3d-clipdrop-bgremove-api": TRI3D_clipdrop_bgremove_api,
+    "tri3d-pose-proportion":TRI3DPoseProportion
 }
 
 VERSION = "2.3"
@@ -2133,5 +2254,6 @@ NODE_DISPLAY_NAME_MAPPINGS = {
     "tri3d-recolor-mask-RGB_space": "Recolor mask RGB space" + " v" + VERSION,
     "tri3d--image-mask-2-box": "Extract box from image" + " v" + VERSION,
     "tri3d-image-mask-box-2-image": "Stitch box to image" + " v" + VERSION,
-    "tri3d-clipdrop-bgremove-api": "RemBG ClipDrop" + " v" + VERSION
+    "tri3d-clipdrop-bgremove-api": "RemBG ClipDrop" + " v" + VERSION,
+    "tri3d-pose-proportion":"Get pose proportion" + "v" + VERSION
 }

dwpose/comfy_utils.py

@@ -35,7 +35,7 @@ def draw_bodypose(canvas: np.ndarray, keypoints: list) -> np.ndarray:
         keypoint1 = keypoints[k1_index - 1]
         keypoint2 = keypoints[k2_index - 1]
 
-        if -1 in keypoint1 or -1 in keypoint2:
+        if any(i < 0 for i in keypoint1) or any(i < 0 for i in keypoint2):
             continue
 
         Y = np.array([keypoint1[0], keypoint2[0]]) 
@@ -46,8 +46,9 @@ def draw_bodypose(canvas: np.ndarray, keypoints: list) -> np.ndarray:
         angle = math.degrees(math.atan2(X[0] - X[1], Y[0] - Y[1]))
         polygon = cv2.ellipse2Poly((int(mY), int(mX)), (int(length / 2), stickwidth), int(angle), 0, 360, 1)
         cv2.fillConvexPoly(canvas, polygon, [int(float(c) * 0.6) for c in color])
+    
     for i, (keypoint, color) in enumerate(zip(keypoints, colors)):
-        if -1 in keypoint:
+        if any(i < 0 for i in keypoint1):
             continue
         
         x, y = keypoint[0], keypoint[1]
@@ -106,25 +107,38 @@ def rotate(src_keypoints, dest_keypoints, point1_idx, point2_idx):
     
     return dest_keypoints
 
-def scale(src_keypoints, dest_keypoints, ref_point1_idx, ref_point2_idx, point1_idx, point2_idx):
-    ref_x1,ref_y1 = src_keypoints[ref_point1_idx]
-    ref_x2,ref_y2 = src_keypoints[ref_point2_idx]
-    ref_x3,ref_y3 = dest_keypoints[ref_point1_idx]
-    ref_x4,ref_y4 = dest_keypoints[ref_point2_idx]
+def scale(src_keypoints, dest_keypoints, ref_point1_idx, ref_point2_idx, point1_idx, point2_idx, ref_torso=False):
+
+    if ref_torso == True:
+        ref_x1,ref_y1 = src_keypoints[1]
+        ref_x3,ref_y3 = dest_keypoints[1]
+
+        ref_x2 = int((src_keypoints[8][0] + src_keypoints[11][0]) / 2)
+        ref_y2 = int((src_keypoints[8][1] + src_keypoints[11][1]) / 2)
+        ref_x4 = int((dest_keypoints[8][0] + dest_keypoints[11][0]) / 2)
+        ref_y4 = int((dest_keypoints[8][1] + dest_keypoints[11][1]) / 2)
+    else:
+        ref_x1,ref_y1 = src_keypoints[ref_point1_idx]
+        ref_x2,ref_y2 = src_keypoints[ref_point2_idx]
+        ref_x3,ref_y3 = dest_keypoints[ref_point1_idx]
+        ref_x4,ref_y4 = dest_keypoints[ref_point2_idx]
     
     x1,y1 = src_keypoints[point1_idx]
     x2,y2 = src_keypoints[point2_idx]
     x3,y3 = dest_keypoints[point1_idx]
     x4,y4 = dest_keypoints[point2_idx]
     
-    src_ref_len = np.linalg.norm(np.array([ref_x1, ref_y1]) - np.array([ref_x2, ref_y2]))             #src ref part distance
-    dest_ref_len = np.linalg.norm(np.array([ref_x3, ref_y3]) - np.array([ref_x4, ref_y4]))          #dest ref part distance
+    # src_ref_len = np.linalg.norm(np.array([ref_x1, ref_y1]) - np.array([ref_x2, ref_y2]))             #src ref part distance
+    # dest_ref_len = np.linalg.norm(np.array([ref_x3, ref_y3]) - np.array([ref_x4, ref_y4]))          #dest ref part distance
 
-    src_targ_len = np.linalg.norm(np.array([x1, y1]) - np.array([x2, y2]))             #src targ part distance
-    dest_targ_len = np.linalg.norm(np.array([x3, y3]) - np.array([x4,y4]))          #dest targ part distance
+    # src_targ_len = np.linalg.norm(np.array([x1, y1]) - np.array([x2, y2]))             #src targ part distance
+    # dest_targ_len = np.linalg.norm(np.array([x3, y3]) - np.array([x4,y4]))          #dest targ part distance
 
-    src_targ_ref_ratio = src_targ_len / src_ref_len                  #src targ to ref ratio
-    dest_targ_ref_ratio = dest_targ_len / dest_ref_len            #dest targ to ref ratio
+    # src_targ_ref_ratio = src_targ_len / src_ref_len                  #src targ to ref ratio
+    # dest_targ_ref_ratio = dest_targ_len / dest_ref_len            #dest targ to ref ratio
+
+    src_targ_ref_ratio = get_ratio(ref_x1, ref_y1, ref_x2, ref_y2, x1, y1, x2, y2)                #src targ to ref ratio
+    dest_targ_ref_ratio = get_ratio(ref_x3, ref_y3, ref_x4, ref_y4, x3, y3, x4, y4)            #dest targ to ref ratio
 
     scale = src_targ_ref_ratio / dest_targ_ref_ratio
 
@@ -135,6 +149,11 @@ def scale(src_keypoints, dest_keypoints, ref_point1_idx, ref_point2_idx, point1_
     
     return dest_keypoints
 
+def get_ratio(ref_x1, ref_y1, ref_x2, ref_y2, x1, y1, x2, y2):
+    ref_len = np.linalg.norm(np.array([ref_x1, ref_y1]) - np.array([ref_x2, ref_y2]))             #ref body part length
+    targ_len = np.linalg.norm(np.array([x1, y1]) - np.array([x2, y2]))             #targ body part length
+    return targ_len / ref_len
+
 def scale_hand(src_keypoints, dest_keypoints, ref_point1_idx, ref_point2_idx, starting_idx):
     
     """
@@ -239,7 +258,7 @@ def get_torso_angles(keypoints):
     a,b = keypoints[5],keypoints[1]
     angle_radians = math.tan((a[1]-b[1])/(a[0]-b[0]))
     rs_angle = abs(math.degrees(angle_radians))
-    
+
     #getting bisector of torso and getting angle with y_axis
     a,b,c = keypoints[8], keypoints[1], keypoints[11]