diff --git a/nerfstudio/process_data/process_data_utils.py b/nerfstudio/process_data/process_data_utils.py
index b5b2391a090..fb2914711bf 100644
--- a/nerfstudio/process_data/process_data_utils.py
+++ b/nerfstudio/process_data/process_data_utils.py
@@ -15,6 +15,7 @@
 """Helper utils for processing data into the nerfstudio format."""
 
 import math
+import random
 import re
 import shutil
 import sys
@@ -126,6 +127,7 @@ def convert_video_to_images(
     verbose: bool = False,
     image_prefix: str = "frame_",
     keep_image_dir: bool = False,
+    random_seed: Optional[int] = None,
 ) -> Tuple[List[str], int]:
     """Converts a video into a sequence of images.
 
@@ -178,8 +180,6 @@ def convert_video_to_images(
             start_y = crop_factor[0]
             crop_cmd = f"crop=w=iw*{width}:h=ih*{height}:x=iw*{start_x}:y=ih*{start_y},"
 
-        spacing = num_frames // num_frames_target
-
         downscale_chains = [f"[t{i}]scale=iw/{2**i}:ih/{2**i}[out{i}]" for i in range(num_downscales + 1)]
         downscale_dirs = [Path(str(image_dir) + (f"_{2**i}" if i > 0 else "")) for i in range(num_downscales + 1)]
         downscale_paths = [downscale_dirs[i] / f"{image_prefix}%05d.png" for i in range(num_downscales + 1)]
@@ -196,7 +196,14 @@ def convert_video_to_images(
 
         ffmpeg_cmd += " -vsync vfr"
 
-        if spacing > 1:
+        # Evenly distribute frame selection if random seed does not exist
+        spacing = num_frames // num_frames_target
+        if random_seed:
+            random.seed(random_seed)
+            frame_indices = sorted(random.sample(range(num_frames), num_frames_target))
+            select_cmd = "select='" + "+".join(["eq(n\\," + str(idx) + ")" for idx in frame_indices]) + "',setpts=N/TB,"
+            CONSOLE.print(f"Extracting {num_frames_target} frames using seed-based random selection.")
+        elif spacing > 1:
             CONSOLE.print("Number of frames to extract:", math.ceil(num_frames / spacing))
             select_cmd = f"thumbnail={spacing},setpts=N/TB,"
         else:
diff --git a/nerfstudio/process_data/video_to_nerfstudio_dataset.py b/nerfstudio/process_data/video_to_nerfstudio_dataset.py
index af17e7d6b6b..f7415bb6522 100644
--- a/nerfstudio/process_data/video_to_nerfstudio_dataset.py
+++ b/nerfstudio/process_data/video_to_nerfstudio_dataset.py
@@ -41,6 +41,10 @@ class VideoToNerfstudioDataset(ColmapConverterToNerfstudioDataset):
     """Feature matching method to use. Vocab tree is recommended for a balance of speed
     and accuracy. Exhaustive is slower but more accurate. Sequential is faster but
     should only be used for videos."""
+    random_seed: Optional[int] = None
+    """Random seed to select video frames"""
+    eval_random_seed: Optional[int] = None
+    """Random seed to select video frames for eval set"""
 
     def main(self) -> None:
         """Process video into a nerfstudio dataset."""
@@ -59,6 +63,7 @@ def main(self) -> None:
                 num_downscales=0,
                 crop_factor=(0.0, 0.0, 0.0, 0.0),
                 verbose=self.verbose,
+                random_seed=self.random_seed,
             )
         else:
             # If we're not dealing with equirects we can downscale in one step.
@@ -71,6 +76,7 @@ def main(self) -> None:
                 verbose=self.verbose,
                 image_prefix="frame_train_" if self.eval_data is not None else "frame_",
                 keep_image_dir=False,
+                random_seed=self.random_seed,
             )
             if self.eval_data is not None:
                 summary_log_eval, num_extracted_frames_eval = process_data_utils.convert_video_to_images(
@@ -82,6 +88,7 @@ def main(self) -> None:
                     verbose=self.verbose,
                     image_prefix="frame_eval_",
                     keep_image_dir=True,
+                    random_seed=self.eval_random_seed,
                 )
                 summary_log += summary_log_eval
                 num_extracted_frames += num_extracted_frames_eval