Project import generated by Copybara.

GitOrigin-RevId: f7d09ed033907b893638a8eb4148efa11c0f09a6

mediapipe/modules/face_landmark/BUILD

@@ -28,12 +28,10 @@ mediapipe_simple_subgraph(
     deps = [
         "//mediapipe/calculators/core:gate_calculator",
         "//mediapipe/calculators/core:split_vector_calculator",
-        "//mediapipe/calculators/image:image_cropping_calculator",
-        "//mediapipe/calculators/image:image_transformation_calculator",
-        "//mediapipe/calculators/tflite:tflite_converter_calculator",
-        "//mediapipe/calculators/tflite:tflite_inference_calculator",
-        "//mediapipe/calculators/tflite:tflite_tensors_to_floats_calculator",
-        "//mediapipe/calculators/tflite:tflite_tensors_to_landmarks_calculator",
+        "//mediapipe/calculators/tensor:image_to_tensor_calculator",
+        "//mediapipe/calculators/tensor:inference_calculator",
+        "//mediapipe/calculators/tensor:tensors_to_floats_calculator",
+        "//mediapipe/calculators/tensor:tensors_to_landmarks_calculator",
         "//mediapipe/calculators/util:landmark_projection_calculator",
         "//mediapipe/calculators/util:thresholding_calculator",
     ],
@@ -46,12 +44,10 @@ mediapipe_simple_subgraph(
     deps = [
         "//mediapipe/calculators/core:gate_calculator",
         "//mediapipe/calculators/core:split_vector_calculator",
-        "//mediapipe/calculators/image:image_cropping_calculator",
-        "//mediapipe/calculators/image:image_transformation_calculator",
-        "//mediapipe/calculators/tflite:tflite_converter_calculator",
-        "//mediapipe/calculators/tflite:tflite_inference_calculator",
-        "//mediapipe/calculators/tflite:tflite_tensors_to_floats_calculator",
-        "//mediapipe/calculators/tflite:tflite_tensors_to_landmarks_calculator",
+        "//mediapipe/calculators/tensor:image_to_tensor_calculator",
+        "//mediapipe/calculators/tensor:inference_calculator",
+        "//mediapipe/calculators/tensor:tensors_to_floats_calculator",
+        "//mediapipe/calculators/tensor:tensors_to_landmarks_calculator",
         "//mediapipe/calculators/util:landmark_projection_calculator",
         "//mediapipe/calculators/util:thresholding_calculator",
     ],
@@ -74,6 +70,7 @@ mediapipe_simple_subgraph(
         "//mediapipe/calculators/image:image_properties_calculator",
         "//mediapipe/calculators/util:association_norm_rect_calculator",
         "//mediapipe/calculators/util:collection_has_min_size_calculator",
+        "//mediapipe/calculators/util:logic_calculator",
         "//mediapipe/modules/face_detection:face_detection_front_cpu",
     ],
 )
@@ -95,6 +92,7 @@ mediapipe_simple_subgraph(
         "//mediapipe/calculators/image:image_properties_calculator",
         "//mediapipe/calculators/util:association_norm_rect_calculator",
         "//mediapipe/calculators/util:collection_has_min_size_calculator",
+        "//mediapipe/calculators/util:logic_calculator",
         "//mediapipe/modules/face_detection:face_detection_front_gpu",
     ],
 )

mediapipe/modules/face_landmark/face_landmark_cpu.pbtxt

@@ -28,51 +28,33 @@ input_stream: "ROI:roi"
 # the absence of this packet so that they don't wait for it unnecessarily.
 output_stream: "LANDMARKS:face_landmarks"
 
-# Crops the input image to the region of interest.
-node {
-  calculator: "ImageCroppingCalculator"
+# Transforms the input image into a 192x192 tensor.
+node: {
+  calculator: "ImageToTensorCalculator"
   input_stream: "IMAGE:image"
   input_stream: "NORM_RECT:roi"
-  output_stream: "IMAGE:face_region"
+  output_stream: "TENSORS:input_tensors"
   options: {
-    [mediapipe.ImageCroppingCalculatorOptions.ext] {
-      border_mode: BORDER_REPLICATE
+    [mediapipe.ImageToTensorCalculatorOptions.ext] {
+      output_tensor_width: 192
+      output_tensor_height: 192
+      output_tensor_float_range {
+        min: 0.0
+        max: 1.0
+      }
     }
   }
 }
 
-# Transforms the input image on CPU to a 192x192 image. To scale the input
-# image, the scale_mode option is set to FIT to preserve the aspect ratio,
-# resulting in potential letterboxing in the transformed image.
-node: {
-  calculator: "ImageTransformationCalculator"
-  input_stream: "IMAGE:face_region"
-  output_stream: "IMAGE:transformed_face_region"
-  options: {
-    [mediapipe.ImageTransformationCalculatorOptions.ext] {
-      output_width: 192
-      output_height: 192
-    }
-  }
-}
-
-# Converts the transformed input image on CPU into an image tensor stored as a
-# TfLiteTensor.
-node {
-  calculator: "TfLiteConverterCalculator"
-  input_stream: "IMAGE:transformed_face_region"
-  output_stream: "TENSORS:input_tensor"
-}
-
 # Runs a TensorFlow Lite model on CPU that takes an image tensor and outputs a
 # vector of tensors representing, for instance, detection boxes/keypoints and
 # scores.
 node {
-  calculator: "TfLiteInferenceCalculator"
-  input_stream: "TENSORS:input_tensor"
+  calculator: "InferenceCalculator"
+  input_stream: "TENSORS:input_tensors"
   output_stream: "TENSORS:output_tensors"
   options: {
-    [mediapipe.TfLiteInferenceCalculatorOptions.ext] {
+    [mediapipe.InferenceCalculatorOptions.ext] {
       model_path: "mediapipe/modules/face_landmark/face_landmark.tflite"
       delegate { xnnpack {} }
     }
@@ -81,7 +63,7 @@ node {
 
 # Splits a vector of tensors into multiple vectors.
 node {
-  calculator: "SplitTfLiteTensorVectorCalculator"
+  calculator: "SplitTensorVectorCalculator"
   input_stream: "output_tensors"
   output_stream: "landmark_tensors"
   output_stream: "face_flag_tensor"
@@ -96,7 +78,7 @@ node {
 # Converts the face-flag tensor into a float that represents the confidence
 # score of face presence.
 node {
-  calculator: "TfLiteTensorsToFloatsCalculator"
+  calculator: "TensorsToFloatsCalculator"
   input_stream: "TENSORS:face_flag_tensor"
   output_stream: "FLOAT:face_presence_score"
 }
@@ -125,11 +107,11 @@ node {
 # Decodes the landmark tensors into a vector of lanmarks, where the landmark
 # coordinates are normalized by the size of the input image to the model.
 node {
-  calculator: "TfLiteTensorsToLandmarksCalculator"
+  calculator: "TensorsToLandmarksCalculator"
   input_stream: "TENSORS:ensured_landmark_tensors"
   output_stream: "NORM_LANDMARKS:landmarks"
   options: {
-    [mediapipe.TfLiteTensorsToLandmarksCalculatorOptions.ext] {
+    [mediapipe.TensorsToLandmarksCalculatorOptions.ext] {
       num_landmarks: 468
       input_image_width: 192
       input_image_height: 192

mediapipe/modules/face_landmark/face_landmark_front_cpu.pbtxt

@@ -26,6 +26,10 @@ input_stream: "IMAGE:image"
 # Max number of faces to detect/track. (int)
 input_side_packet: "NUM_FACES:num_faces"
 
+# Whether face detection can be skipped when face regions can already be
+# approximated from face landmarks on the previous frame.
+input_side_packet: "CAN_SKIP_DETECTION:can_skip_detection"
+
 # Collection of detected/predicted faces, each represented as a list of 468 face
 # landmarks. (std::vector<NormalizedLandmarkList>)
 # NOTE: there will not be an output packet in the LANDMARKS stream for this
@@ -54,12 +58,22 @@ node {
 }
 
 # Drops the incoming image if FaceLandmarkCpu was able to identify face presence
-# in the previous image. Otherwise, passes the incoming image through to trigger
-# a new round of face detection in FaceDetectionFrontCpu.
+# in the previous image and skipping face detection is enabled. Otherwise,
+# passes the incoming image through to trigger a new round of face detection
+# in FaceDetectionFrontCpu.
+node {
+  calculator: "LogicCalculator"
+  options: {
+    [mediapipe.LogicCalculatorOptions.ext] { op: AND }
+  }
+  input_side_packet: "can_skip_detection"
+  input_stream: "prev_has_enough_faces"
+  output_stream: "skip_detection"
+}
 node {
   calculator: "GateCalculator"
   input_stream: "image"
-  input_stream: "DISALLOW:prev_has_enough_faces"
+  input_stream: "DISALLOW:skip_detection"
   output_stream: "gated_image"
   options: {
     [mediapipe.GateCalculatorOptions.ext] {
@@ -67,6 +81,12 @@ node {
     }
   }
 }
+node {
+  calculator: "GateCalculator"
+  input_stream: "prev_face_rects_from_landmarks"
+  input_stream: "ALLOW:skip_detection"
+  output_stream: "gated_prev_face_rects_from_landmarks"
+}
 
 # Detects faces.
 node {
@@ -129,7 +149,7 @@ node {
 # overlapping regions based on the specified min_similarity_threshold.
 node {
   calculator: "AssociationNormRectCalculator"
-  input_stream: "prev_face_rects_from_landmarks"
+  input_stream: "gated_prev_face_rects_from_landmarks"
   input_stream: "face_rects_from_detections"
   output_stream: "face_rects"
   options: {

mediapipe/modules/face_landmark/face_landmark_front_gpu.pbtxt

@@ -26,6 +26,10 @@ input_stream: "IMAGE:image"
 # Max number of faces to detect/track. (int)
 input_side_packet: "NUM_FACES:num_faces"
 
+# Whether face detection can be skipped when face regions can already be
+# approximated from face landmarks on the previous frame.
+input_side_packet: "CAN_SKIP_DETECTION:can_skip_detection"
+
 # Collection of detected/predicted faces, each represented as a list of 468 face
 # landmarks. (std::vector<NormalizedLandmarkList>)
 # NOTE: there will not be an output packet in the LANDMARKS stream for this
@@ -54,12 +58,22 @@ node {
 }
 
 # Drops the incoming image if FaceLandmarkGpu was able to identify face presence
-# in the previous image. Otherwise, passes the incoming image through to trigger
-# a new round of face detection in FaceDetectionFrontGpu.
+# in the previous image and skipping face detection is enabled. Otherwise,
+# passes the incoming image through to trigger a new round of face detection
+# in FaceDetectionFrontGpu.
+node {
+  calculator: "LogicCalculator"
+  options: {
+    [mediapipe.LogicCalculatorOptions.ext] { op: AND }
+  }
+  input_side_packet: "can_skip_detection"
+  input_stream: "prev_has_enough_faces"
+  output_stream: "skip_detection"
+}
 node {
   calculator: "GateCalculator"
   input_stream: "image"
-  input_stream: "DISALLOW:prev_has_enough_faces"
+  input_stream: "DISALLOW:skip_detection"
   output_stream: "gated_image"
   options: {
     [mediapipe.GateCalculatorOptions.ext] {
@@ -67,6 +81,12 @@ node {
     }
   }
 }
+node {
+  calculator: "GateCalculator"
+  input_stream: "prev_face_rects_from_landmarks"
+  input_stream: "ALLOW:skip_detection"
+  output_stream: "gated_prev_face_rects_from_landmarks"
+}
 
 # Detects faces.
 node {
@@ -129,7 +149,7 @@ node {
 # overlapping regions based on the specified min_similarity_threshold.
 node {
   calculator: "AssociationNormRectCalculator"
-  input_stream: "prev_face_rects_from_landmarks"
+  input_stream: "gated_prev_face_rects_from_landmarks"
   input_stream: "face_rects_from_detections"
   output_stream: "face_rects"
   options: {

mediapipe/modules/face_landmark/face_landmark_gpu.pbtxt

@@ -1,5 +1,5 @@
-# MediaPipe graph to detect/predict face landmarks. (GPU input, and inference is
-# executed on GPU.)
+# MediaPipe graph to detect/predict face landmarks. (CPU input, and inference is
+# executed on CPU.)
 #
 # It is required that "face_landmark.tflite" is available at
 # "mediapipe/modules/face_landmark/face_landmark.tflite"
@@ -28,51 +28,34 @@ input_stream: "ROI:roi"
 # the absence of this packet so that they don't wait for it unnecessarily.
 output_stream: "LANDMARKS:face_landmarks"
 
-# Crops the input image to the given region of interest.
-node {
-  calculator: "ImageCroppingCalculator"
+# Transforms the input image into a 192x192 tensor.
+node: {
+  calculator: "ImageToTensorCalculator"
   input_stream: "IMAGE_GPU:image"
   input_stream: "NORM_RECT:roi"
-  output_stream: "IMAGE_GPU:face_region"
+  output_stream: "TENSORS:input_tensors"
   options: {
-    [mediapipe.ImageCroppingCalculatorOptions.ext] {
-      border_mode: BORDER_REPLICATE
+    [mediapipe.ImageToTensorCalculatorOptions.ext] {
+      output_tensor_width: 192
+      output_tensor_height: 192
+      output_tensor_float_range {
+        min: 0.0
+        max: 1.0
+      }
+      gpu_origin: TOP_LEFT
     }
   }
 }
 
-# Transforms the input image on GPU to a 192x192 image. To scale the input
-# image, the scale_mode option is set to FIT to preserve the aspect ratio,
-# resulting in potential letterboxing in the transformed image.
-node: {
-  calculator: "ImageTransformationCalculator"
-  input_stream: "IMAGE_GPU:face_region"
-  output_stream: "IMAGE_GPU:transformed_face_region"
-  options: {
-    [mediapipe.ImageTransformationCalculatorOptions.ext] {
-      output_width: 192
-      output_height: 192
-    }
-  }
-}
-
-# Converts the transformed input image on GPU into an image tensor stored as a
-# TfLiteTensor.
-node {
-  calculator: "TfLiteConverterCalculator"
-  input_stream: "IMAGE_GPU:transformed_face_region"
-  output_stream: "TENSORS_GPU:input_tensor"
-}
-
 # Runs a TensorFlow Lite model on GPU that takes an image tensor and outputs a
 # vector of GPU tensors representing, for instance, detection boxes/keypoints
 # and scores.
 node {
-  calculator: "TfLiteInferenceCalculator"
-  input_stream: "TENSORS_GPU:input_tensor"
+  calculator: "InferenceCalculator"
+  input_stream: "TENSORS:input_tensors"
   output_stream: "TENSORS:output_tensors"
   options: {
-    [mediapipe.TfLiteInferenceCalculatorOptions.ext] {
+    [mediapipe.InferenceCalculatorOptions.ext] {
       model_path: "mediapipe/modules/face_landmark/face_landmark.tflite"
     }
   }
@@ -80,7 +63,7 @@ node {
 
 # Splits a vector of tensors into multiple vectors.
 node {
-  calculator: "SplitTfLiteTensorVectorCalculator"
+  calculator: "SplitTensorVectorCalculator"
   input_stream: "output_tensors"
   output_stream: "landmark_tensors"
   output_stream: "face_flag_tensor"
@@ -95,7 +78,7 @@ node {
 # Converts the face-flag tensor into a float that represents the confidence
 # score of face presence.
 node {
-  calculator: "TfLiteTensorsToFloatsCalculator"
+  calculator: "TensorsToFloatsCalculator"
   input_stream: "TENSORS:face_flag_tensor"
   output_stream: "FLOAT:face_presence_score"
 }
@@ -124,11 +107,11 @@ node {
 # Decodes the landmark tensors into a vector of lanmarks, where the landmark
 # coordinates are normalized by the size of the input image to the model.
 node {
-  calculator: "TfLiteTensorsToLandmarksCalculator"
+  calculator: "TensorsToLandmarksCalculator"
   input_stream: "TENSORS:ensured_landmark_tensors"
   output_stream: "NORM_LANDMARKS:landmarks"
   options: {
-    [mediapipe.TfLiteTensorsToLandmarksCalculatorOptions.ext] {
+    [mediapipe.TensorsToLandmarksCalculatorOptions.ext] {
       num_landmarks: 468
       input_image_width: 192
       input_image_height: 192