DETAILED ACTION
Notice of Pre-AIA or AIA Status
1. The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
2. This is in response to the applicant response filed on 11/25/2025. In the applicant’s response, claims 1, and 3 are amended; claim 2 was cancelled. Accordingly, claims 1 and 3-8 are pending and being examined. Claim 1 is independent form.
Claim Rejections - 35 USC § 102
3. In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
4. The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale or otherwise available to the public before the effective filing date of the claimed invention.
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
5. Claims 1, 3, 4, and 8 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Ge et al (“Efficient Low-Resolution Face Recognition via Bridge Distillation”, IEEE, 2020, hereinafter “Ge”).
Regarding claim 1, Ge discloses a feature vector transfer method for improving face recognition performance of a low-quality image performed by at least one processor (the bridge distillation approach for low-resolution face recognition; see abstract and fig.2), the feature vector transfer method comprising:
training a high-quality face recognition network for recognizing a human face based on a high-quality image including the human face; extracting a first feature vector associated with the high-quality image from the high-quality face recognition network (see the top part of fig.2 and Sec. III-A, 4) Formulation, lines 15-17: “1) cross-dataset distillation adapts the pre-learned knowledge in high-resolution faces from private domain to public domain and distils it to compact features”);
transferring the extracted first feature vector onto a low-quality face recognition network for recognizing the human face based on a low-quality image including the human face; and training the low-quality face recognition network using the transferred first feature vector (see the bottom part of fig.2 and Sec. III-A, 4) Formulation, lines 18-21: “2) resolution-adapted distillation transfers the knowledge from high-resolution public faces to their low-resolution versions by taking the adapted features as supervision signals to train the student model [using low-resolution face images].”), wherein the training of the low-quality face recognition network comprises: extracting a second feature vector from the low-quality face recognition network; and training the low-quality face recognition network so that a direction of the second feature vector becomes similar to a direction of the first feature vector by using knowledge distillation (see Sec. III-A-3) Student model: “[t]he student model has much simpler design than that of the teacher. By training the student to mimic the teacher’s behavior on the public dataset IS, the model’s complexity is largely reduced. In practical settings, our aim is to obtain an optimized student model Ms that works well on the low-resolution target dataset IT ”. See Eq (6), in the left col, on page 6903, which is updated to make the feature vector Mls (.) generated by the student model as similar as possible to the compact adapted feature vector Fa (.) transferred from the teacher stream. It should be noticed that, two vectors are similar means both the magnitudes and the directions of them are similar. See the paragraph right below Eq(3), lines 7-13: “[a]s a result, the adapted features [i.e., Fa (.)] mimic the impact of the discriminative selection of the original high-resolution features, but with greatly reduced dimensions. After training the adaptation module Ma, we discard the softmax layer and only retain the adapted and reduced features
f
a
I
=
F
a
(
[
.
]
)
as supervision for training the student model.” Therefore, the feature vector Mls (.) is the “second feature vector from the low-quality face recognition network” while the compact adapted feature vector Fa (.) is “the first feature vector” generated from the high-resolution dataset).
Regarding claim 3, Ge discloses the feature vector transfer method of claim 2, wherein the training of the low- quality face recognition network so that the direction of the second feature vector becomes similar to the direction of the first feature vector comprises training the low-quality face recognition network using a sum of a face recognition loss and a distillation loss in the low-quality face recognition network (wherein the student model is trained by Eqs.(4)-(6) on the low-resolution images to minimize both the face classification loss and the feature regression loss. See Sec. III-C).
Regarding claim 4, Ge discloses the feature vector transfer method of claim 1, further comprising: acquiring the high-quality image including the human face (e.g., see “public dataset (High-Res)” in fig.2); performing downsampling on the acquired high-quality image (e.g., see “public dataset (Low-Res)” in fig.2); performing blur processing on the downsampled image; and generating the low quality image by changing a size of the blurred image to a size corresponding to the high quality image (e.g., see Sec. IV, para.2: “UMDFaces serves as the public dataset IS, which contains 367,888 images in 8,419 subjects. To generate high-resolution public images, faces are normalized into 224 × 224 and 112×112 sizes for learning the VGGFace2 and CosFace adaptation
module. Low-resolution public face images are achieved by randomly perturbing the localized facial landmarks for 16 times, normalizing to various resolutions of p × p and degrading to approximate the distribution of target faces (e.g., blurring, changing illumination, etc).”).
Regarding claim 8, claim 8 is an inherent variation of claim 1, thus it is interpreted and rejected for the reasons set forth in the rejection of claim 1.
Claim Rejections - 35 USC § 103
6. The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102 of this title, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
7. Claim 5-7 are rejected under 35 U.S.C. 103 as being unpatentable over Ge in view of Ullah et al (“Robust Face Recognition Method for Occluded and Low-Resolution Images”, 2019, hereinafter “Ullah”).
Regarding claim 5, Ge does not disclose, extracting a first attention map associated with the high-quality image from the trained high-quality face recognition network; and transferring the extracted first attention map onto the low-quality face recognition network for recognizing the human face based on the low-quality image including the human face, as recited in the claim. However, this technique is well known and widely used in the field of face recognition. As evidence, in the same field of endeavor, Ullah teaches a face recognition method which may extract “68 definite [robust and landmark] points exist around the outer edge of both eyes, the top of the chin, and the eyebrows’ inner edge”. See Sec. III-A, para.1, lines 6-10. It would have been obvious to one of ordinary skill in the art before the effective filling date of the claimed invention was made to incorporate the teachings of Ullah into the teachings of Ge and further extract the robust landmarks from facial images taught by Ullah for low-resolution face recognition method taught by Ge. Suggestion or motivation for doing so would have been to “handles occlusion and low-resolution on two publicly available databases in near real-time” as taught by Ullah, see Sec. I, bullet one and Abstract. Therefore, the claim is unpatentable over Ge in view of Ullah.
Regarding claim 6, the combination of Ge and Ullah discloses the feature vector transfer method of claim 5, wherein the training of the low-quality face recognition network further comprises: extracting a second attention map from the low-quality face recognition network; and training the low-quality face recognition network so that the second attention map becomes similar to the first attention map by using knowledge distillation (these features are obvious for one of ordinary skill in the art based the teachings of the combination of Ge and Ullah. E.g., Ge, see fig.2 and Eqs(4)-(6)).
Regarding claim 7, the combination of Ge and Ullah discloses the feature vector transfer method of claim 5, wherein the high-quality face recognition network comprises a plurality of blocks for extracting the first feature vector of the high-quality image and a plurality of attention modules for extracting the first attention map (these features are obvious for one of ordinary skill in the art based the teachings of the combination of Ge and Ullah. E.g., Ge, see fig.2 and Eqs(4)-(6)).
Response to Arguments
8. Applicant’s arguments, with respects to claim 1 filed on 11/25/2025, have been fully considered but they are not persuasive.
On page 7 of applicant’s response, applicant argues:
Ge, however, does not disclose or suggest "wherein the training of the low-quality face recognition network comprises: extracting a second feature vector from the low-quality face recognition network; and training the low-quality face recognition network so that a direction of the second feature vector becomes similar to a direction of the first feature vector by using knowledge distillation," as recited in amended claim 1...
Furthermore, since Ge does not disclose or suggest extracting the claimed "second feature vector" and training the student model so that "a direction of the second feature vector becomes similar to a direction of the first feature vector," Ge also does not disclose or suggest any directional feature distillations such as the angular similarity loss, normalized embeddings, and cosine-based distillation.
The examiner respectfully disagrees with the applicant’s arguments. As explained in the rejection of the claim, Ge clearly discloses: “By training the student to mimic the teacher’s behavior on the public dataset IS, the model’s complexity is largely reduced. In practical settings, our aim is to obtain an optimized student model Ms that works well on the low-resolution target dataset IT ”. “As a result, the adapted features mimic the impact of the discriminative selection of the original high-resolution features, but with greatly reduced dimensions. After training the adaptation module Ma, we discard the softmax layer and only retain the adapted and reduced features
f
a
I
=
F
a
(
[
.
]
)
as supervision for training the student model.” As shown by Eq(6), the student model is updated to make the feature vector Mls (.) generated by the student model using the low-resolution target dataset as similar as possible to the compact adapted feature vector Fa (.) transferred from the teacher stream which is based on the high-resolution dataset. Therefore, the arguments are not persuasive.
9. In view of the above reasons, examiner maintains rejections.
Conclusion
10. THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any extension fee pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
11. Any inquiry concerning this communication or earlier communications from the examiner should be directed to RUIPING LI whose telephone number is (571)270-3376. The examiner can normally be reached 8:30am--5:30pm.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, HENOK SHIFERAW can be reached on (571)272-4637. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit https://patentcenter.uspto.gov; https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center, and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/RUIPING LI/Primary Examiner, Ph.D., Art Unit 2676
Prosecution Insights