add stfpm infer (#2087)

paddlex/inference/components/transforms/image/common.py

@@ -19,6 +19,7 @@ from copy import deepcopy
 
 import numpy as np
 import cv2
+from skimage import measure, morphology
 
 from .....utils.cache import CACHE_DIR
 from ....utils.io import ImageReader, ImageWriter
@@ -539,3 +540,30 @@ class ToCHWImage(BaseComponent):
         """apply"""
         img = img.transpose((2, 0, 1))
         return {"img": img}
+
+
+class Map_to_mask(BaseComponent):
+    """Map_to_mask"""
+    INPUT_KEYS = "pred"
+    OUTPUT_KEYS = "pred"
+    DEAULT_INPUTS = {"pred": "pred"}
+    DEAULT_OUTPUTS = {"pred": "pred"}
+
+    def apply(self, pred):
+        """apply"""
+        # from skimage import measure, morphology
+        # import cv2
+        # from PIL import Image
+        # import numpy as np
+        # import imageio
+
+        score_map = pred[0]
+        thred = 0.01
+        mask = score_map[0]
+        mask[mask > thred] = 255
+        mask[mask <= thred] = 0
+        kernel = morphology.disk(4)
+        mask = morphology.opening(mask, kernel)
+        mask = mask.astype(np.uint8)
+
+        return {"pred": mask[None, :, :]}

paddlex/inference/models/__init__.py

@@ -29,6 +29,7 @@ from .ts_fc import TSFcPredictor
 from .ts_ad import TSAdPredictor
 from .ts_cls import TSClsPredictor
 from .image_unwarping import WarpPredictor
+from .anomaly_detection import UadPredictor
 
 
 def _create_hp_predictor(

paddlex/inference/models/anomaly_detection.py

@@ -0,0 +1,87 @@
+# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#    http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import numpy as np
+
+from ...utils.func_register import FuncRegister
+from ...modules.anomaly_detection.model_list import MODELS
+from ..components import *
+from ..results import SegResult
+from ..utils.process_hook import batchable_method
+from .base import CVPredictor
+from ..components.transforms.image.common import Map_to_mask 
+
+class UadPredictor(CVPredictor):
+
+    entities = MODELS
+
+    _FUNC_MAP = {}
+    register = FuncRegister(_FUNC_MAP)
+
+    def _build_components(self):
+        self._add_component(ReadImage(format="RGB"))
+        for cfg in self.config["Deploy"]["transforms"]:
+            tf_key = cfg["type"]
+            func = self._FUNC_MAP.get(tf_key)
+            cfg.pop("type")
+            args = cfg
+            op = func(self, **args) if args else func(self)
+            self._add_component(op)
+        self._add_component(ToCHWImage())
+        predictor = ImagePredictor(
+            model_dir=self.model_dir,
+            model_prefix=self.MODEL_FILE_PREFIX,
+            option=self.pp_option,
+        )
+        self._add_component(("Predictor", predictor))
+        self._add_component(Map_to_mask())
+
+    @register("Resize")
+    def build_resize(
+        self, target_size, keep_ratio=False, size_divisor=None, interp="LINEAR"
+    ):
+        assert target_size
+        op = Resize(
+            target_size=target_size,
+            keep_ratio=keep_ratio,
+            size_divisor=size_divisor,
+            interp=interp,
+        )
+        return op
+
+    @register("ResizeByLong")
+    def build_resizebylong(self, long_size):
+        assert long_size
+        return ResizeByLong(
+            target_long_edge=long_size, size_divisor=size_divisor, interp=interp
+        )
+
+    @register("ResizeByShort")
+    def build_resizebylong(self, short_size):
+        assert short_size
+        return ResizeByLong(
+            target_long_edge=short_size, size_divisor=size_divisor, interp=interp
+        )
+
+    @register("Normalize")
+    def build_normalize(
+        self,
+        mean=0.5,
+        std=0.5,
+    ):
+        return Normalize(mean=mean, std=std)
+
+    def _pack_res(self, single):
+        keys = ["img_path", "pred"]
+        return SegResult({key: single[key] for key in keys})