add pose filter

4 files changed, 252 insertions, 78 deletions

diff --git a/megapixels/app/processors/face_pose.py b/megapixels/app/processors/face_pose.py
index 67ac685d..f2548b32 100644
--- a/megapixels/app/processors/face_pose.py
+++ b/megapixels/app/processors/face_pose.py
@@ -22,89 +22,83 @@ class FacePoseDLIB:
   def __init__(self):
     pass
 
-  def pose(self, landmarks, dim): 
-    '''Calculates pose
-    '''
-    degrees = compute_pose_degrees(landmarks, dim)    
-    return degrees
 
+  def pose(self, landmarks, dim, project_points=False):
+    # computes pose using 6 / 68 points from dlib face landmarks
+    # based on learnopencv.com and 
+    # https://github.com/jerryhouuu/Face-Yaw-Roll-Pitch-from-Pose-Estimation-using-OpenCV/
+    # NB: not as accurate as MTCNN, see @jerryhouuu for ideas
+    
+    pose_points_idx = (30, 8, 36, 45, 48, 54)
+    axis = np.float32([[500,0,0], [0,500,0], [0,0,500]])
+    
+    # 3D model points.
+    model_points = np.array([
+        (0.0, 0.0, 0.0),             # Nose tip
+        (0.0, -330.0, -65.0),        # Chin
+        (-225.0, 170.0, -135.0),     # Left eye left corner
+        (225.0, 170.0, -135.0),      # Right eye right corne
+        (-150.0, -150.0, -125.0),    # Left Mouth corner
+        (150.0, -150.0, -125.0)      # Right mouth corner
+    ])
+    
+    # Assuming no lens distortion
+    dist_coeffs = np.zeros((4,1))
 
-# -----------------------------------------------------------
-# utilities
-# -----------------------------------------------------------
+    # find 6 pose points
+    pose_points = []
+    for j, idx in enumerate(pose_points_idx):
+      pt = landmarks[idx]
+      pose_points.append((pt[0], pt[1]))
+    pose_points = np.array(pose_points, dtype='double')  # convert to double
+    
+    # create camera matrix
+    focal_length = dim[0]
+    center = (dim[0]/2, dim[1]/2)
+    cam_mat = np.array(
+        [[focal_length, 0, center[0]],
+        [0, focal_length, center[1]],
+        [0, 1, 1]], dtype = "double")
+    
+    # solve PnP for rotation and translation
+    (success, rot_vec, tran_vec) = cv.solvePnP(model_points, pose_points, 
+                                               cam_mat, dist_coeffs, 
+                                               flags=cv.SOLVEPNP_ITERATIVE)
 
-def compute_pose_degrees(landmarks, dim):
-  # computes pose using 6 / 68 points from dlib face landmarks
-  # based on learnopencv.com and 
-  # https://github.com/jerryhouuu/Face-Yaw-Roll-Pitch-from-Pose-Estimation-using-OpenCV/
-  # NB: not as accurate as MTCNN, see @jerryhouuu for ideas
-  
-  pose_points_idx = (30, 8, 36, 45, 48, 54)
-  axis = np.float32([[500,0,0], [0,500,0], [0,0,500]])
-  
-  # 3D model points.
-  model_points = np.array([
-      (0.0, 0.0, 0.0),             # Nose tip
-      (0.0, -330.0, -65.0),        # Chin
-      (-225.0, 170.0, -135.0),     # Left eye left corner
-      (225.0, 170.0, -135.0),      # Right eye right corne
-      (-150.0, -150.0, -125.0),    # Left Mouth corner
-      (150.0, -150.0, -125.0)      # Right mouth corner
-  ])
-  
-  # Assuming no lens distortion
-  dist_coeffs = np.zeros((4,1))
+    result = {}
 
-  # find 6 pose points
-  pose_points = []
-  for j, idx in enumerate(pose_points_idx):
-    pt = landmarks[idx]
-    pose_points.append((pt[0], pt[1]))
-  pose_points = np.array(pose_points, dtype='double')  # convert to double
-  
-  # create camera matrix
-  focal_length = dim[0]
-  center = (dim[0]/2, dim[1]/2)
-  cam_mat = np.array(
-      [[focal_length, 0, center[0]],
-      [0, focal_length, center[1]],
-      [0, 1, 1]], dtype = "double")
-  
-  # solve PnP for rotation and translation
-  (success, rot_vec, tran_vec) = cv.solvePnP(model_points, pose_points, 
-                                             cam_mat, dist_coeffs, 
-                                             flags=cv.SOLVEPNP_ITERATIVE)
+    # project points
+    if project_points:
+      pts_im, jac = cv.projectPoints(axis, rot_vec, tran_vec, cam_mat, dist_coeffs)
+      pts_model, jac2 = cv.projectPoints(model_points, rot_vec, tran_vec, cam_mat, dist_coeffs)
+      result['points_model'] = pts_model
+      result['points_image'] = pts_im
+      result['point_nose'] = tuple(landmarks[pose_points_idx[0]])
 
-  # project points
-  #pts_im, jac = cv.projectPoints(axis, rot_vec, tran_vec, cam_mat, dist_coeffs)
-  #pts_model, jac2 = cv.projectPoints(model_points, rot_vec, tran_vec, cam_mat, dist_coeffs)
-  rvec_matrix = cv.Rodrigues(rot_vec)[0]
-  
-  # convert to degrees
-  proj_matrix = np.hstack((rvec_matrix, tran_vec))
-  eulerAngles = cv.decomposeProjectionMatrix(proj_matrix)[6] 
-  pitch, yaw, roll = [math.radians(x) for x in eulerAngles]
-  pitch = math.degrees(math.asin(math.sin(pitch)))
-  roll = -math.degrees(math.asin(math.sin(roll)))
-  yaw = math.degrees(math.asin(math.sin(yaw)))
-  degrees = {'pitch': pitch, 'roll': roll, 'yaw': yaw}
-  
-  # add nose point
-  #pt_nose = tuple(landmarks[pose_points_idx[0]])
-  return degrees
-  #return pts_im, pts_model, degrees, pt_nose
+    rvec_matrix = cv.Rodrigues(rot_vec)[0]
+    
+    # convert to degrees
+    proj_matrix = np.hstack((rvec_matrix, tran_vec))
+    eulerAngles = cv.decomposeProjectionMatrix(proj_matrix)[6] 
+    pitch, yaw, roll = [math.radians(x) for x in eulerAngles]
+    pitch = math.degrees(math.asin(math.sin(pitch)))
+    roll = -math.degrees(math.asin(math.sin(roll)))
+    yaw = math.degrees(math.asin(math.sin(yaw)))
+    degrees = {'pitch': pitch, 'roll': roll, 'yaw': yaw}
+    result['degrees'] = degrees
 
+    return result
 
-def draw_pose(im, pts_im, pts_model, pt_nose):
-  cv.line(im, pt_nose, tuple(pts_im[1].ravel()), (0,255,0), 3) #GREEN
-  cv.line(im, pt_nose, tuple(pts_im[0].ravel()), (255,0,), 3) #BLUE
-  cv.line(im, pt_nose, tuple(pts_im[2].ravel()), (0,0,255), 3) #RED
-  return im
 
+  def draw_pose(self, im, pts_im, pts_model, pt_nose):
+    cv.line(im, pt_nose, tuple(pts_im[1].ravel()), (0,255,0), 3) #GREEN
+    cv.line(im, pt_nose, tuple(pts_im[0].ravel()), (255,0,), 3) #BLUE
+    cv.line(im, pt_nose, tuple(pts_im[2].ravel()), (0,0,255), 3) #RED
 
-def draw_degrees(im, degrees, color=(0,255,0)):
-  for i, item in enumerate(degrees.items()):
-    k, v = item
-    t = '{}: {:.2f}'.format(k, v)
-    origin = (10, 30 + (25 * i))
-    cv.putText(im, t, origin, cv.FONT_HERSHEY_SIMPLEX, 0.5, color, thickness=2, lineType=2)
\ No newline at end of file
+
+  def draw_degrees(self, im, degrees, color=(0,255,0)):
+    for i, item in enumerate(degrees.items()):
+      k, v = item
+      t = '{}: {:.2f}'.format(k, v)
+      origin = (10, 30 + (25 * i))
+      cv.putText(im, t, origin, cv.FONT_HERSHEY_SIMPLEX, 0.5, color, thickness=2, lineType=2)
\ No newline at end of file
diff --git a/megapixels/app/settings/app_cfg.py b/megapixels/app/settings/app_cfg.py
index 4c540231..7406caad 100644
--- a/megapixels/app/settings/app_cfg.py
+++ b/megapixels/app/settings/app_cfg.py
@@ -8,6 +8,10 @@ import cv2 as cv
 from app.settings import types
 from app.utils import click_utils
 
+# -----------------------------------------------------------------------------
+# Metadata type names
+# -----------------------------------------------------------------------------
+
 
 # -----------------------------------------------------------------------------
 # Enun lists used for custom Click Params
@@ -16,6 +20,8 @@ from app.utils import click_utils
 FaceDetectNetVar = click_utils.ParamVar(types.FaceDetectNet)
 HaarCascadeVar = click_utils.ParamVar(types.HaarCascade)
 LogLevelVar = click_utils.ParamVar(types.LogLevel)
+MetadataVar = click_utils.ParamVar(types.Metadata)
+DatasetVar = click_utils.ParamVar(types.Dataset)
 
 # # data_store
 DATA_STORE = '/data_store_hdd/'
@@ -23,6 +29,7 @@ DATA_STORE_NAS = '/data_store_nas/'
 DATA_STORE_HDD = '/data_store_hdd/'
 DATA_STORE_SSD = '/data_store_ssd/'
 DIR_DATASETS = join(DATA_STORE,'datasets')
+DIR_DATSET_NAS = join(DIR_DATASETS, 'people')
 DIR_APPS = join(DATA_STORE,'apps')
 DIR_APP = join(DIR_APPS,'megapixels')
 DIR_MODELS = join(DIR_APP,'models')
diff --git a/megapixels/app/settings/paths.py b/megapixels/app/settings/paths.py
new file mode 100644
index 00000000..bc1333ba
--- /dev/null
+++ b/megapixels/app/settings/paths.py
@@ -0,0 +1,163 @@
+import os
+from os.path import join
+import logging
+
+from vframe.settings import vframe_cfg as vcfg
+from vframe.settings import types
+
+class Paths:
+  
+  # class properties
+  MAPPINGS_DATE = vcfg.SUGARCUBE_DATES[0]
+  DIR_APP_VFRAME = 'apps/vframe/'
+  DIR_APP_SA = 'apps/syrianarchive'
+  DIR_MODELS_VFRAME = join(DIR_APP_VFRAME, 'models')
+  DIR_DARKNET = join(DIR_MODELS_VFRAME, 'darknet/pjreddie')
+  DIR_DARKNET_VFRAME = join(DIR_MODELS_VFRAME, 'darknet/vframe')
+  DIR_MEDIA = join(DIR_APP_SA, 'media')
+  DIR_METADATA = join(DIR_APP_SA, 'metadata')
+  DIR_RECORDS = join(DIR_APP_SA, 'records')
+  DIR_REPORTS = join(DIR_APP_SA, 'reports')
+
+
+  def __init__(self):
+    pass
+
+  @classmethod
+  def DataStorePath(cls, data_store=types.DataStore.HDD):
+    return '/data_store_{}'.format(data_store.name.lower())
+
+  # -------------------------------------------------------------------------------
+  # Darknet Paths
+
+  @classmethod
+  def darknet_classes(cls, data_store=types.DataStore.HDD, opt_net=types.DetectorNet.COCO):
+    if opt_net == types.DetectorNet.COCO:
+      fp = join(cls.DIR_DARKNET, 'data', 'coco.names')
+    elif opt_net == types.DetectorNet.COCO_SPP:
+      fp = join(cls.DIR_DARKNET, 'data', 'coco.names')
+    elif opt_net == types.DetectorNet.VOC:
+      fp = join(cls.DIR_DARKNET, 'data', 'voc.names')
+    elif opt_net == types.DetectorNet.OPENIMAGES:
+      fp = join(cls.DIR_DARKNET, 'data', 'openimages.names')
+    elif opt_net == types.DetectorNet.SUBMUNITION:
+      fp = join(cls.DIR_DARKNET_VFRAME, 'munitions_09b', 'classes.txt')
+    return join(cls.DataStorePath(data_store), fp)
+
+  @classmethod
+  def darknet_data(cls, data_store=types.DataStore.HDD, opt_net=types.DetectorNet.COCO, as_bytes=True):
+    if opt_net == types.DetectorNet.COCO:
+      fp = join(cls.DIR_DARKNET, 'cfg', 'coco.data')
+    elif opt_net == types.DetectorNet.COCO_SPP:
+      fp = join(cls.DIR_DARKNET, 'cfg', 'coco.data')
+    elif opt_net == types.DetectorNet.VOC:
+      fp = join(cls.DIR_DARKNET, 'cfg', 'voc.data')
+    elif opt_net == types.DetectorNet.OPENIMAGES:
+      fp = join(cls.DIR_DARKNET, 'cfg', 'openimages.data')
+    elif opt_net == types.DetectorNet.SUBMUNITION:
+      fp = join(cls.DIR_DARKNET_VFRAME, 'munitions_09b', 'meta.data')
+    fp = join(cls.DataStorePath(data_store), fp)
+    if as_bytes:
+      return bytes(fp, encoding="utf-8")
+    else:
+      return fp
+
+
+  @classmethod
+  def darknet_cfg(cls, data_store=types.DataStore.HDD, opt_net=types.DetectorNet.COCO, as_bytes=True):
+    if opt_net == types.DetectorNet.COCO:
+      fp = join(cls.DIR_DARKNET, 'cfg', 'yolov3.cfg')
+    elif opt_net == types.DetectorNet.COCO_SPP:
+      fp = join(cls.DIR_DARKNET, 'cfg', 'yolov3-spp.cfg')
+    elif opt_net == types.DetectorNet.VOC:
+      fp = join(cls.DIR_DARKNET, 'cfg', 'yolov3-voc.cfg')
+    elif opt_net == types.DetectorNet.OPENIMAGES:
+      fp = join(cls.DIR_DARKNET, 'cfg', 'yolov3-openimages.cfg')
+    elif opt_net == types.DetectorNet.SUBMUNITION:
+      fp = join(cls.DIR_DARKNET_VFRAME, 'munitions_09b', 'yolov3.cfg')
+    fp = join(cls.DataStorePath(data_store), fp)
+    if as_bytes:
+      return bytes(fp, encoding="utf-8")
+    else:
+      return fp
+
+  @classmethod
+  def darknet_weights(cls, data_store=types.DataStore.HDD, opt_net=types.DetectorNet.COCO, as_bytes=True):
+    if opt_net == types.DetectorNet.COCO:
+      fp = join(cls.DIR_DARKNET, 'weights', 'yolov3.weights')
+    elif opt_net == types.DetectorNet.COCO_SPP:
+      fp = join(cls.DIR_DARKNET, 'weights', 'yolov3-spp.weights')
+    elif opt_net == types.DetectorNet.VOC:
+      fp = join(cls.DIR_DARKNET, 'weights', 'yolov3-voc.weights')
+    elif opt_net == types.DetectorNet.OPENIMAGES:
+      fp = join(cls.DIR_DARKNET, 'weights', 'yolov3-openimages.weights')
+    elif opt_net == types.DetectorNet.SUBMUNITION:
+      fp = join(cls.DIR_DARKNET_VFRAME, 'munitions_09b/weights', 'yolov3_40000.weights')
+    fp = join(cls.DataStorePath(data_store), fp)
+    if as_bytes:
+      return bytes(fp, encoding="utf-8")
+    else:
+      return fp
+
+  # -------------------------------------------------------------------------------
+  # Metadata Paths
+
+  @classmethod
+  def mapping_index(cls, opt_date, data_store=types.DataStore.HDD, verified=types.Verified.VERIFIED, 
+    file_format=types.FileExt.PKL):
+    """Returns filepath to a mapping file. Mapping files are the original Suguarcube mapping data"""
+    fname = 'index.pkl' if file_format == types.FileExt.PKL else 'index.json'
+    # data_store = 'data_store_{}'.format(data_store.name.lower())
+    date_str = opt_date.name.lower()
+    fp = join(cls.DataStorePath(data_store), cls.DIR_METADATA, 'mapping', date_str, verified.name.lower(), fname)
+    return fp
+
+  @classmethod
+  def media_record_index(cls, data_store=types.DataStore.HDD, verified=types.Verified.VERIFIED, 
+    file_format=types.FileExt.PKL):
+    """Returns filepath to a mapping file. Mapping files are the original Suguarcube mapping data"""
+    fname = 'index.pkl' if file_format == types.FileExt.PKL else 'index.json'
+    metadata_type = types.Metadata.MEDIA_RECORD.name.lower()
+    fp = join(cls.DataStorePath(data_store), cls.DIR_METADATA, metadata_type, verified.name.lower(), fname)
+    return fp
+
+  @classmethod
+  def metadata_index(cls, metadata_type, data_store=types.DataStore.HDD, 
+    verified=types.Verified.VERIFIED, file_format=types.FileExt.PKL):
+    """Uses key from enum to get folder name and construct filepath"""
+    fname = 'index.pkl' if file_format == types.FileExt.PKL else 'index.json'
+    fp = join(cls.DataStorePath(data_store), cls.DIR_METADATA, metadata_type.name.lower(), 
+      verified.name.lower(), fname)
+    return fp
+
+  @classmethod
+  def metadata_dir(cls, metadata_type, data_store=types.DataStore.HDD, verified=types.Verified.VERIFIED):
+    """Uses key from enum to get folder name and construct filepath"""
+    fp = join(cls.DataStorePath(data_store), cls.DIR_METADATA, metadata_type.name.lower(), 
+      verified.name.lower())
+    return fp
+
+  @classmethod
+  def metadata_tree_dir(cls, metadata_type, data_store=types.DataStore.HDD):
+    """Uses key from enum to get folder name and construct filepath"""
+    fp = join(cls.DataStorePath(data_store), cls.DIR_METADATA, metadata_type.name.lower())
+    return fp
+
+  @classmethod
+  def media_dir(cls, media_type, data_store=types.DataStore.HDD, verified=types.Verified.VERIFIED):
+    """Returns the directory path to a media directory"""
+    fp = join(cls.DataStorePath(data_store), cls.DIR_MEDIA, media_type.name.lower(), verified.name.lower())
+    return fp
+
+  # @classmethod
+  # def keyframe(cls, dir_media, idx, image_size=types.ImageSize.MEDIUM):
+  #   """Returns path to keyframe image using supplied cls.media directory"""
+  #   idx = str(idx).zfill(vcfg.ZERO_PADDING)
+  #   size_label = vcfg.IMAGE_SIZE_LABELS[image_size]
+  #   fp = join(dir_media, sha256_tree, sha256, idx, size_label, 'index.jpg')
+  #   return fp
+
+  @classmethod
+  def dnn(cls):
+    """Returns configurations for available DNNs"""
+    pass
\ No newline at end of file
diff --git a/megapixels/app/settings/types.py b/megapixels/app/settings/types.py
index e9107803..7157436d 100644
--- a/megapixels/app/settings/types.py
+++ b/megapixels/app/settings/types.py
@@ -7,7 +7,6 @@ def find_type(name, enum_type):
   return None
 
 
-
 class FaceDetectNet(Enum):
   """Scene text detector networks"""
   HAAR, DLIB_CNN, DLIB_HOG, CVDNN, MTCNN = range(5)
@@ -31,3 +30,14 @@ class HaarCascade(Enum):
 class LogLevel(Enum):
   """Loger vebosity"""
   DEBUG, INFO, WARN, ERROR, CRITICAL = range(5)
+
+
+# ---------------------------------------------------------------------
+# Metadata types
+# --------------------------------------------------------------------
+
+class Metadata(Enum):
+  IDENTITIES, POSES, ROIS, FILE_META, SHAS, UUIDS, FACE_VECTORS = range(7)
+
+class Dataset(Enum):
+  LFW, VGG_FACE2 = range(2)